Distributed Learning Environment in Multicultural Context A Symposium

国际自动化与计算杂志.英文版.1.Improved Exponential Stability Criteria for Uncertain Neutral System with Nonlinear Parameter PerturbationsFang Qiu，Ban-Tong Cui2.Robust Active Suspension Design Subject to Vehicle Inertial Parameter VariationsHai-Ping Du，Nong Zhang3.Delay-dependent Non-fragile H∞ Filtering for Uncertain Fuzzy Systems Based on Switching Fuzzy Model and Piecewise Lyapunov FunctionZhi-Le Xia，Jun-Min Li，Jiang-Rong Li4.Observer-based Adaptive Iterative Learning Control for Nonlinear Systems with Time-varying DelaysWei-Sheng Chen，Rui-Hong Li，Jing Li5.H∞ Output Feedback Control for Stochastic Systems with Mode-dependent Time-varying Delays and Markovian Jump ParametersXu-Dong Zhao，Qing-Shuang Zeng6.Delay and Its Time-derivative Dependent Robust Stability of Uncertain Neutral Systems with Saturating ActuatorsFatima El Haoussi，El Houssaine Tissir7.Parallel Fuzzy P+Fuzzy I+Fuzzy D Controller:Design and Performance EvaluationVineet Kumar，A.P.Mittal8.Observers for Descriptor Systems with Slope-restricted NonlinearitiesLin-Na Zhou，Chun-Yu Yang，Qing-Ling Zhang9.Parameterized Solution to a Class of Sylvester MatrixEquationsYu-Peng Qiao，Hong-Sheng Qi，Dai-Zhan Cheng10.Indirect Adaptive Fuzzy and Impulsive Control of Nonlinear SystemsHai-Bo Jiang11.Robust Fuzzy Tracking Control for Nonlinear Networked Control Systems with Integral Quadratic ConstraintsZhi-Sheng Chen，Yong He，Min Wu12.A Power-and Coverage-aware Clustering Scheme for Wireless Sensor NetworksLiang Xue，Xin-Ping Guan，Zhi-Xin Liu，Qing-Chao Zheng13.Guaranteed Cost Active Fault-tolerant Control of Networked Control System with Packet Dropout and Transmission DelayXiao-Yuan Luo，Mei-Jie Shang，Cai-Lian Chen，Xin-Ping Guanparison of Two Novel MRAS Based Strategies for Identifying Parameters in Permanent Magnet Synchronous MotorsKan Liu，Qiao Zhang，Zi-Qiang Zhu，Jing Zhang，An-Wen Shen，Paul Stewart15.Modeling and Analysis of Scheduling for Distributed Real-time Embedded SystemsHai-Tao Zhang，Gui-Fang Wu16.Passive Steganalysis Based on Higher Order Image Statistics of Curvelet TransformS.Geetha，Siva S.Sivatha Sindhu，N.Kamaraj17.Movement Invariants-based Algorithm for Medical Image Tilt CorrectionMei-Sen Pan，Jing-Tian Tang，Xiao-Li Yang18.Target Tracking and Obstacle Avoidance for Multi-agent SystemsJing Yan，Xin-Ping Guan，Fu-Xiao Tan19.Automatic Generation of Optimally Rigid Formations Using Decentralized MethodsRui Ren，Yu-Yan Zhang，Xiao-Yuan Luo，Shao-Bao Li20.Semi-blind Adaptive Beamforming for High-throughput Quadrature Amplitude Modulation SystemsSheng Chen，Wang Yao，Lajos Hanzo21.Throughput Analysis of IEEE 802.11 Multirate WLANs with Collision Aware Rate Adaptation AlgorithmDhanasekaran Senthilkumar，A. Krishnan22.Innovative Product Design Based on Customer Requirement Weight Calculation ModelChen-Guang Guo，Yong-Xian Liu，Shou-Ming Hou，Wei Wang23.A Service Composition Approach Based on Sequence Mining for Migrating E-learning Legacy System to SOAZhuo Zhang，Dong-Dai Zhou，Hong-Ji Yang，Shao-Chun Zhong24.Modeling of Agile Intelligent Manufacturing-oriented Production Scheduling SystemZhong-Qi Sheng，Chang-Ping Tang，Ci-Xing Lv25.Estimation of Reliability and Cost Relationship for Architecture-based SoftwareHui Guan，Wei-Ru Chen，Ning Huang，Hong-Ji Yang1.A Computer-aided Design System for Framed-mould in Autoclave ProcessingTian-Guo Jin，Feng-Yang Bi2.Wear State Recognition of Drills Based on K-means Cluster and Radial Basis Function Neural NetworkXu Yang3.The Knee Joint Design and Control of Above-knee Intelligent Bionic Leg Based on Magneto-rheological DamperHua-Long Xie，Ze-Zhong Liang，Fei Li，Li-Xin Guo4.Modeling of Pneumatic Muscle with Shape Memory Alloy and Braided SleeveBin-Rui Wang，Ying-Lian Jin，Dong Wei5.Extended Object Model for Product Configuration DesignZhi-Wei Xu，Ze-Zhong Liang，Zhong-Qi Sheng6.Analysis of Sheet Metal Extrusion Process Using Finite Element MethodXin-Cun Zhuang，Hua Xiang，Zhen Zhao7.Implementation of Enterprises' Interoperation Based on OntologyXiao-Feng Di，Yu-Shun Fan8.Path Planning Approach in Unknown EnvironmentTing-Kai Wang，Quan Dang，Pei-Yuan Pan9.Sliding Mode Variable Structure Control for Visual Servoing SystemFei Li，Hua-Long Xie10.Correlation of Direct Piezoelectric Effect on EAPap under Ambient FactorsLi-Jie Zhao，Chang-Ping Tang，Peng Gong11.XML-based Data Processing in Network Supported Collaborative DesignQi Wang，Zhong-Wei Ren，Zhong-Feng Guo12.Production Management Modelling Based on MASLi He，Zheng-Hao Wang，Ke-Long Zhang13.Experimental Tests of Autonomous Ground Vehicles with PreviewCunjia Liu，Wen-Hua Chen，John Andrews14.Modelling and Remote Control of an ExcavatorYang Liu，Mohammad Shahidul Hasan，Hong-Nian Yu15.TOPSIS with Belief Structure for Group Belief Multiple Criteria Decision MakingJiang Jiang，Ying-Wu Chen，Da-Wei Tang，Yu-Wang Chen16.Video Analysis Based on Volumetric Event DetectionJing Wang，Zhi-Jie Xu17.Improving Decision Tree Performance by Exception HandlingAppavu Alias Balamurugan Subramanian，S.Pramala，B.Rajalakshmi，Ramasamy Rajaram18.Robustness Analysis of Discrete-time Indirect Model Reference Adaptive Control with Normalized Adaptive LawsQing-Zheng Gao，Xue-Jun Xie19.A Novel Lifecycle Model for Web-based Application Development in Small and Medium EnterprisesWei Huang，Ru Li，Carsten Maple，Hong-Ji Yang，David Foskett，Vince Cleaver20.Design of a Two-dimensional Recursive Filter Using the Bees AlgorithmD. T. Pham，Ebubekir Ko(c)21.Designing Genetic Regulatory Networks Using Fuzzy Petri Nets ApproachRaed I. Hamed，Syed I. Ahson，Rafat Parveen1.State of the Art and Emerging Trends in Operations and Maintenance of Offshore Oil and Gas Production Facilities: Some Experiences and ObservationsJayantha P.Liyanage2.Statistical Safety Analysis of Maintenance Management Process of Excavator UnitsLjubisa Papic，Milorad Pantelic，Joseph Aronov，Ajit Kumar Verma3.Improving Energy and Power Efficiency Using NComputing and Approaches for Predicting Reliability of Complex Computing SystemsHoang Pham，Hoang Pham Jr.4.Running Temperature and Mechanical Stability of Grease as Maintenance Parameters of Railway BearingsJan Lundberg，Aditya Parida，Peter S(o)derholm5.Subsea Maintenance Service Delivery: Mapping Factors Influencing Scheduled Service DurationEfosa Emmanuel Uyiomendo，Tore Markeset6.A Systemic Approach to Integrated E-maintenance of Large Engineering PlantsAjit Kumar Verma，A.Srividya，P.G.Ramesh7.Authentication and Access Control in RFID Based Logistics-customs Clearance Service PlatformHui-Fang Deng，Wen Deng，Han Li，Hong-Ji Yang8.Evolutionary Trajectory Planning for an Industrial RobotR.Saravanan，S.Ramabalan，C.Balamurugan，A.Subash9.Improved Exponential Stability Criteria for Recurrent Neural Networks with Time-varying Discrete and Distributed DelaysYuan-Yuan Wu，Tao Li，Yu-Qiang Wu10.An Improved Approach to Delay-dependent Robust Stabilization for Uncertain Singular Time-delay SystemsXin Sun，Qing-Ling Zhang，Chun-Yu Yang，Zhan Su，Yong-Yun Shao11.Robust Stability of Nonlinear Plants with a Non-symmetric Prandtl-Ishlinskii Hysteresis ModelChang-An Jiang，Ming-Cong Deng，Akira Inoue12.Stability Analysis of Discrete-time Systems with Additive Time-varying DelaysXian-Ming Tang，Jin-Shou Yu13.Delay-dependent Stability Analysis for Markovian Jump Systems with Interval Time-varying-delaysXu-Dong Zhao，Qing-Shuang Zeng14.H∞ Synchronization of Chaotic Systems via Delayed Feedback ControlLi Sheng，Hui-Zhong Yang15.Adaptive Fuzzy Observer Backstepping Control for a Class of Uncertain Nonlinear Systems with Unknown Time-delayShao-Cheng Tong，Ning Sheng16.Simulation-based Optimal Design of α-β-γ-δ FilterChun-Mu Wu，Paul P.Lin，Zhen-Yu Han，Shu-Rong Li17.Independent Cycle Time Assignment for Min-max SystemsWen-De Chen，Yue-Gang Tao，Hong-Nian Yu1.An Assessment Tool for Land Reuse with Artificial Intelligence MethodDieter D. Genske，Dongbin Huang，Ariane Ruff2.Interpolation of Images Using Discrete Wavelet Transform to Simulate Image Resizing as in Human VisionRohini S. Asamwar，Kishor M. Bhurchandi，Abhay S. Gandhi3.Watermarking of Digital Images in Frequency DomainSami E. I. Baba，Lala Z. Krikor，Thawar Arif，Zyad Shaaban4.An Effective Image Retrieval Mechanism Using Family-based Spatial Consistency Filtration with Object RegionJing Sun，Ying-Jie Xing5.Robust Object Tracking under Appearance Change ConditionsQi-Cong Wang，Yuan-Hao Gong，Chen-Hui Yang，Cui-Hua Li6.A Visual Attention Model for Robot Object TrackingJin-Kui Chu，Rong-Hua Li，Qing-Ying Li，Hong-Qing Wang7.SVM-based Identification and Un-calibrated Visual Servoing for Micro-manipulationXin-Han Huang，Xiang-Jin Zeng，Min Wang8.Action Control of Soccer Robots Based on Simulated Human IntelligenceTie-Jun Li，Gui-Qiang Chen，Gui-Fang Shao9.Emotional Gait Generation for a Humanoid RobotLun Xie，Zhi-Liang Wang，Wei Wang，Guo-Chen Yu10.Cultural Algorithm for Minimization of Binary Decision Diagram and Its Application in Crosstalk Fault DetectionZhong-Liang Pan，Ling Chen，Guang-Zhao Zhang11.A Novel Fuzzy Direct Torque Control System for Three-level Inverter-fed Induction MachineShu-Xi Liu，Ming-Yu Wang，Yu-Guang Chen，Shan Li12.Statistic Learning-based Defect Detection for Twill FabricsLi-Wei Han，De Xu13.Nonsaturation Throughput Enhancement of IEEE 802.11b Distributed Coordination Function for Heterogeneous Traffic under Noisy EnvironmentDhanasekaran Senthilkumar，A. Krishnan14.Structure and Dynamics of Artificial Regulatory Networks Evolved by Segmental Duplication and Divergence ModelXiang-Hong Lin，Tian-Wen Zhang15.Random Fuzzy Chance-constrained Programming Based on Adaptive Chaos Quantum Honey Bee Algorithm and Robustness AnalysisHan Xue，Xun Li，Hong-Xu Ma16.A Bit-level Text Compression Scheme Based on the ACW AlgorithmHussein A1-Bahadili，Shakir M. Hussain17.A Note on an Economic Lot-sizing Problem with Perishable Inventory and Economies of Scale Costs:Approximation Solutions and Worst Case AnalysisQing-Guo Bai，Yu-Zhong Zhang，Guang-Long Dong1.Virtual Reality: A State-of-the-Art SurveyNing-Ning Zhou，Yu-Long Deng2.Real-time Virtual Environment Signal Extraction and DenoisingUsing Programmable Graphics HardwareYang Su，Zhi-Jie Xu，Xiang-Qian Jiang3.Effective Virtual Reality Based Building Navigation Using Dynamic Loading and Path OptimizationQing-Jin Peng，Xiu-Mei Kang，Ting-Ting Zhao4.The Skin Deformation of a 3D Virtual HumanXiao-Jing Zhou，Zheng-Xu Zhao5.Technology for Simulating Crowd Evacuation BehaviorsWen-Hu Qin，Guo-Hui Su，Xiao-Na Li6.Research on Modelling Digital Paper-cut PreservationXiao-Fen Wang，Ying-Rui Liu，Wen-Sheng Zhang7.On Problems of Multicomponent System Maintenance ModellingTomasz Nowakowski，Sylwia Werbinka8.Soft Sensing Modelling Based on Optimal Selection of Secondary Variables and Its ApplicationQi Li，Cheng Shao9.Adaptive Fuzzy Dynamic Surface Control for Uncertain Nonlinear SystemsXiao-Yuan Luo，Zhi-Hao Zhu，Xin-Ping Guan10.Output Feedback for Stochastic Nonlinear Systems with Unmeasurable Inverse DynamicsXin Yu，Na Duan11.Kalman Filtering with Partial Markovian Packet LossesBao-Feng Wang，Ge Guo12.A Modified Projection Method for Linear FeasibilityProblemsYi-Ju Wang，Hong-Yu Zhang13.A Neuro-genetic Based Short-term Forecasting Framework for Network Intrusion Prediction SystemSiva S. Sivatha Sindhu，S. Geetha，M. Marikannan，A. Kannan14.New Delay-dependent Global Asymptotic Stability Condition for Hopfield Neural Networks with Time-varying DelaysGuang-Deng Zong，Jia Liu hHTTp://15.Crosscumulants Based Approaches for the Structure Identification of Volterra ModelsHouda Mathlouthi，Kamel Abederrahim，Faouzi Msahli，Gerard Favier1.Coalition Formation in Weighted Simple-majority Games under Proportional Payoff Allocation RulesZhi-Gang Cao，Xiao-Guang Yang2.Stability Analysis for Recurrent Neural Networks with Time-varying DelayYuan-Yuan Wu，Yu-Qiang Wu3.A New Type of Solution Method for the Generalized Linear Complementarity Problem over a Polyhedral ConeHong-Chun Sun，Yan-Liang Dong4.An Improved Control Algorithm for High-order Nonlinear Systems with Unmodelled DynamicsNa Duan，Fu-Nian Hu，Xin Yu5.Controller Design of High Order Nonholonomic System with Nonlinear DriftsXiu-Yun Zheng，Yu-Qiang Wu6.Directional Filter for SAR Images Based on NonsubsampledContourlet Transform and Immune Clonal SelectionXiao-Hui Yang，Li-Cheng Jiao，Deng-Feng Li7.Text Extraction and Enhancement of Binary Images Using Cellular AutomataG. Sahoo，Tapas Kumar，B.L. Rains，C.M. Bhatia8.GH2 Control for Uncertain Discrete-time-delay Fuzzy Systems Based on a Switching Fuzzy Model and Piecewise Lyapunov FunctionZhi-Le Xia，Jun-Min Li9.A New Energy Optimal Control Scheme for a Separately Excited DC Motor Based Incremental Motion DriveMilan A.Sheta，Vivek Agarwal，Paluri S.V.Nataraj10.Nonlinear Backstepping Ship Course ControllerAnna Witkowska，Roman Smierzchalski11.A New Method of Embedded Fourth Order with Four Stages to Study Raster CNN SimulationR. Ponalagusamy，S. Senthilkumar12.A Minimum-energy Path-preserving Topology Control Algorithm for Wireless Sensor NetworksJin-Zhao Lin，Xian Zhou，Yun Li13.Synchronization and Exponential Estimates of Complex Networks with Mixed Time-varying Coupling DelaysYang Dai，YunZe Cai，Xiao-Ming Xu14.Step-coordination Algorithm of Traffic Control Based on Multi-agent SystemHai-Tao Zhang，Fang Yu，Wen Li15.A Research of the Employment Problem on Common Job-seekersand GraduatesBai-Da Qu。

Data SheetKey SpecificationsKey Features•2x2 MU-MIMO with two spatialstreams per radio•Third 2x2 MIMO radio for dedicated RFand WIPS scanning•802.11ac Wave 2 support •Up to 400 Mbps for 2.4 GHz radio •Up to 867 Mbps for 5 GHz radio •Integrated omnidirectional antennas •20/40/80 MHz channel width support •Integrated BLE •2x Gigabit Ethernet port•Full Operational Capacity with 802.3atPoE+•Distributed Data Plane architecture •Zero-touch deployment through automatic cloud activation and configuration •Cloud or on premises management plane options •Operating modes for dedicated access, dedicated security or dual-mode •Support for up to 8 distinct SSIDs per radio •Integrated firewall, traffic shaping, QoS and BYOD controls per SSID •Dynamic RF optimization through smart steering, band steering and optimal channel selection •Application visibility through layer 7 deep packet inspection •Automated device access logging •Patented Marker Packettm technology for rogue AP detection and classification •Wired VLAN monitoring for “No-WiFi” zone enforcement •Third party analytics integration with real-time data transfer •Self-healing wireless mesh networkingTop Performance at the Best PriceArista W-118 is an enterprise-grade 2x2 MU-MIMO tri-radio 802.11ac wall plate access point with dual concurrent 5 GHz and 2.4 GHz radios supporting 802.11a/n/ ac Wave 2, 802.11b/g/n, two spatial streams, and data rates of up to 876 Mbps and 300 Mbps, respectively. It also contains a third 2x2 MIMO 802.11ac radio for dedicated multi-function scanning and a fourth 2.4 GHz Bluetooth Low Energy (BLE) radio.Why Choose the W-118?The W-118 provides best value amongst high-performing, modern wall plate access points designed for cost conscious organizations. Built using the latest 802.11ac Wave 2 chipsets, the W-118 is perfect for medium density environments looking for the high performance and advanced features of current accesspoints without the high cost. Common deployment scenarios include small and medium schools, distributed remote offices, small meeting rooms, and enterprise campuses.The W-118 provides access to advanced access point features like role-based firewalls and application visibility without the high cost typically associated with Wave 2 devices. The W-118 is also a perfect fit for organizations in need of future-ready dedicated security sensorsiBeacon Bluetooth Low Energy SupportThe Arista W-118 supports the iBeacon Bluetooth Low Energy (BLE) standard. BLE is used for proximity based services on mobile devices via an application ecosystem. W-118 can be configured to advertise a unique identifier through iBeacons at a periodic interval.Arista Cloud Managed WiFiThe W-118 is managed by the Arista cloud and leverages a purpose-built cloud architecture to produce enterprise-grade wireless networks for every application required, ensuring high reliability through an approach that is automated, scalable, secure and cost effective.What Really MattersThe future of WiFi requires intelligent, self-reliant access points that support high-performing, highly reliable networks without the need for antiquated controllers. This approach removes the complexity, instability and high costs associated with enterprise WiFi today.Arista W-118AccessThe W-118 creates WiFi networks that require less time and resources to deploy and maintain compared to traditional devices, resulting in significant cost savings.• Plug and play provisioning using either Cloud or On-premise deployments - Arista Access Points take less than two minutes to activate and configure after connecting to the cloud• Support for up to eight individual SSIDs per radio providing maximum flexibility in network design• Network controls like NAT, Firewall and QoS implemented at the Access Point, ensuring faster and more reliable networks• Continuous scanning of all 2.4 GHz and 5 GHz channels by a dedicated 2x2 third radio provides a dynamic, 360 degree view of the RF environment to assist in RF optimization and client handling• Network availability and performance assurance using the third radio as a client to conduct on-demand and scheduled connectivity and performance tests• Smart steering addresses sticky client issues by automatically pushing clients with low data rates to a better access point• Band steering manages channel occupancy, pushing clients to the 5 GHz channel for optimal throughput• Smart load balancing distributes load evenly across neighbouring APs to optimize the use of network resources• Arista Wi-Fi’s distributed data plane architecture continues to serve users and secure the network even if connection with the management plane is interrupted• Interference avoidance from LTE/3G small/macro cells in commonly used TDD/FDD frequency bandsSecurityThe W-118 offers complete visibility and control of the wireless airspace that keeps the integrity of the network in check and actively protects users without manual intervention.•W-118 is equipped with industry leading fully integrated wireless intrusion prevention capabilities•Multifunction third radio provides uninterrupted spectrum scanning or client emulation for always on security coverage alongside dedicated 2.4G/5G client radios.•Arista’s patented Marker PacketsTM help accurately detect rogue access points on any network while minimizing false positives •Third radio used as a dedicated security sensor for 24x7x365 scanning and automated over-the-air (OTA) prevention •Deterministic rogue AP detection and prevention by monitoring all WiFi and non-WiFi VLANs.•Over-the-air and on-the-wire prevention techniques assure automatic and reliable threat prevention to keep unauthorized clientsand rogue APs off the network without impacting authorized connections.•Access Points autonomously scan for wireless threats and enforce security policy even if disconnected from the cloud management plane•VLAN monitoring enables a virtual connection to non-WiFi networks for complete network rogue detection and prevention AnalyticsThe W-118 collects massive amounts of data and supports immersive guest network experiences that develop and reinforce the relationship between them and the brand.• Reports of customer footfall, demographic, loyalty and other analytics provide insightful and actionable information.• Supports proximity marketing programs that trigger when certain devices are present, which includes automatic messaging vis MMSin-browser notifications and real time notifications sent to 3rd party systems that alert to the presence of enrolled devices.Property SpecificationPhysical Dimensions186.4mm X 123.9mm X 25.5mm / 7.3” X 4.9” X 1”Weight .455kg (1 lb)Operating Temperature 0o C – 40o C (32o F – 104o F)Storage Temperature -25o C – 75o C (-13o F – 167o F)MTBF535,205 hr @ 40o C1,081,559 hr @ 25o CHumidity0%-95% non-condensingP ower consuption11.8W (max) / 5.1W (min) / 8.3W (avg)Chipset Qualcomm QCA4019 SOCProcessor RAMQualcomm IPQ4019 717MHz quad core ARMprocessor with 512 MB RAM and 32 MB Flash Physical SpecificationsPort Description Connector Type Speed/ProtocolPower12V 2A5.5mm overalldiameter/2.1mmcenter pin/holePass-throughportThe pass-through port is used to pluga device into another wired port thatis available on the wall where the AP isinstalled. The pass-through port at therear of the device and pass-throughport on the bottom of the device areinternally connected.RJ-45--Ethernet(LAN3/PSE)Gigabit Ethernet port that can be usedfor wired extension for an SSID. Thisport also provides the power for thedevice using the 802.3af standardRJ-4510/100/1000 MbpsGigabit EthernetEthernet(LAN2/LAN1)Gigabit Ethernet port that can be usedfor wired extension for an SSID.RJ-4510/100/1000 MbpsGigabit EthernetReset Reset to factory default settingsPin hole pushbuttonHold down andpower cycle thedevice to resetOperational Specifications Port DescriptionConnectorTypeSpeed/Pro-tocol PassthroughThis is a wired port that facilitatesextension of the wired network after theAP is mounted on the wall. Another devicecan be plugged in to the pass-through porton the bottom of the W-118 device. Thetraffic on the pass-through port does notinterfere with the AP traffic. No policies canbe applied on the pass-through port traffic.RJ-45-WANEnables the connection to wired LANthrough a switch or hub. The device canthen communicate with the server. Thisport also provides the power for the deviceusing the 802.3af standardRJ-4510/100/1000Mbps EthernetPower overEthernetInput Power12V DC 2ANumber of Radios 3 WiFi Radios: One 2.4 GHz and 5 GHz radio each for simultaneous dual band client access. Athird dual-band radio dedicated to non-access smart scanning; WIPS, RF optimization, Remote Troubleshooting, and network assurance functions.1 BLE Radio: A fourth Bluetooth Low Energy radio for proximity based services on mobile devices via an application ecosystem.Max Clients Supported512 clients per radio (dependent upon use cases)MIMO2x2 for 2.4/5GHz RadiosNumber of Spatial Streams 2 for 2.4/5GHz RadiosRF Transmit Power20 dBm per radio chain (max); Actual power for Tx will depend on Country Regulatory Domain Simultaneous MU-MIMO Clients Two 1x1 MU-MIMO clientsUsers in a MU-MIMO group with a2x2 client1Bandwidth Agility YesFrequency Bands 2.4-2.4835 GHz, 4.9-5.0 GHz, 5.15-5.25 GHz (UNII-1), 5.25-5.35 GHz, 5.47-5.6 GHz,5.650-5.725 GHz (UNII-2), 5.725-5.85 GHz (UNII-3)Dynamic Frequency Selection Supported in compliance to all latest amendments from FCC, CE, IC, CB, TELEC, KCC regarding certifications.Frequency, Modulation and Data RatesIEEE 802.11b/g/nFrequency BandScanning TransmissionAll regionsUSA & Canada(FCC/IC)Europe(ETSI) 2400 ~ 2483.5 MHz2400 ~ 2473.5 MHz2400 ~ 2483.5 MHzModulation Type DSSS, OFDMPeak Data Rates Up to 300 Mbps (MCS 0-15)Antenna Integrated modular high efficiency PIFA antenna x4 (peak gain 5.0 dBi)IEEE 802.11a/n/acFrequency Band Scanning TransmissionAll regions USA & Canada(FCC/IC)Europe(ETSI)4.92 ~5.08 GHz5.15 ~ 5.25 GHz 5.25 ~ 5.35 GHz 5.47 ~ 5.725 GHz 5.725 ~ 5.825 GHz 5.15 ~ 5.25 GHz5.25 ~ 5.35 GHz5.725 ~ 5.825 GHz5.15 ~ 5.25 GHz5.25 ~ 5.35 GHz5.47 ~ 5.725 GHzDynamic Frequency Selection DFS and DFS2Modulation Type OFDMPeak Data Rates Up to 867 Mbps (MCS 0-15)Antenna Integrated modular high efficiency PIFA antenna x4 (peak gain 5.0 dBi)Maximum Aggregate Transmit PowerFor 2.4 GHzMCS Index Transmit Power(dBm)802.11b1 Mbps -11 Mbps22802.11g6 Mbps - 48 Mbps2554 Mbps802.11n HT20MCS 0,1,2,3,4,524802.11n HT40MCS 0,1,2,3,4,5 24For 5 GHzMCS Index Transmit Power(dBm)802.11a6 Mbps - 48 Mbps26.802.11n HT20MCS 0,1,2,3,4,526802.11n HT40MCS 0,1,2,3,4,526802.11ac VHT80MCS 0,1,2,3,4,5,6,726Note:The actual transmit power will be the lowest of:• Value specified in the Device Template• Maximum value allowed in the regulatory domain • Maximum power supported by the radioData Sheet Receive SensitivityFor 2.4 GHzMCS Index Receive Sensitivity (dBm)802.11g6 Mbps -9224 Mbps -36 Mbps -48 Mbps -54 Mbps -75802.11n HT20MCS 0, 8 -92MCS 1,9MCS 2,10MCS 3,11MCS 4.12MCS 5,13MCS 6,14MCS 7, 15 -73802.11n HT40MCS 0, 8 -89MCS 1,9MCS 2,10MCS 3,11MCS 4,12MCS 5,13MCS 6,14MCS 7, 15 -71.5 For 5 GHzMCS Index Receive Sensitivity (dBm)802.11a6 Mbps -9024 Mbps36 Mbps48 Mbps54 Mbps -74.5802.11n HT20MCS 0, 8 -90MCS 1,9MCS 2,10MCS 3,11MCS 4,12MCS 5,13MCS 6,14MCS 7,15 -73802.11n HT40MCS 0, 8 -88.5MCS 1,9MCS 2,10MCS 3,11MCS 4,12MCS 5,13MCS 6,14MCS 7, 15 -70For 5 GHzMCS Index Receive Sensitivity (dBm)802.11n VHT20MCS 0 -90MCS 1MCS 2MCS 3MCS 4MCS 5MCS 6MCS 7MCS 8 -69802.11n VHT40MCS 9-65802.11n VHT80MCS 0 -85.5MCS 1MCS 2MCS 3MCS 4MCS 5MCS 6MCS 7MCS 8MCS 9 -61Data Sheet5 GHz2.4 GHzInternal Antenna Radiation Patterns Internal Antenna Radiation Patterns dBi gaindBi gainData SheetHeadquarters5453 Great America Parkway Santa Clara, California 95054408-547-5500Copyright 2020 Arista Networks, Inc. The information contained herein is subject to change without notice. Arista, the Arista logo and EOS are trademarks of Arista Networks. Other product or service names may be trademarks or service marks of others.Support******************408-547-5502866-476-0000Sales****************408-547-5501866-497-0000Ordering Information : Access Point Power Part Number DescriptionAP-W118-SS-5Y W-118 2x2:2 tri radio 802.11ac Wave-2 wall plate access point with internal antennas and 5 year Cognitive Cloud SW SubscriptionAP-W118-SS-3Y W-118 2x2:2 tri radio 802.11ac Wave-2 wall plate access point with internal antennas and 3 year Cognitive Cloud SW SubscriptionAP-W118W-118 2x2:2 tri radio 802.11ac Wave-2 wall plate access point with internal antennas Part Number DescriptionPWR-AP-W4Universal AC power supply for all APs except for C-110PWR-AP-PLUS-NA One port 802.3at PoE+ injector for use with all Access Point models. Includes USA power cord. Not for outdoor use.”PWR-AP-W2Universal AC power supply for C-120, C-130, W-118 and C-100October 1, 2020Regulatory Specifications RF and ElectromagneticCountry CertificationUSA FCC Part 15.247, 15.407EuropeCE EN300.328, EN301.893Countries covered under Europe certification: Austria, Belgium, Cyprus, Denmark, Estonia, Finland, France,Germany, Greece, Hungary, Ireland, Italy, Iceland, Luxembourg, Latvia, Lithuania, Malta, Netherlands, Norway,Poland, Portugal, Spain, Sweden, Slovakia, Slovenia, Switzerland, The Czech Republic, UK.CountryCertificationUSA UL 60950CanadacUL 60950European Union (EU)EN 60950, RoHSSafety*For complete country certification records, please visit the site: https:///en/support/product-certificate AP-W118-R WW-118-R W 2x2:2 tri radio 802.11ac Wave-2 wall plate access point with internal antennas (bundled with Stand, Power supply, Ethernet cable)PWR-AP-W3Non-discountable purchase. Universal AC power supply for W-118, C-120, C-130 and C-100, 12VDC, 2A, Center +, DC Plug 5.5mm*2.1mm*L9.5mm, US UK Euro AU Plugs。

多智能体强化学习的几种BestPractice（草稿阶段，完成度40%）多智能体强化学习的几种Best Practice - vonZooming的文章 - 知乎 https:///p/99120143这里分享一下A Survey and Critique of Multiagent Deep Reinforcement Learning这篇综述里面介绍的多智能体强化学习Best Practice。

这部分内容大部分来自第四章，但是我根据自己的理解加上了其他的内容。

1.改良Experience replay buffer1.1 传统的Single-agent场景之下的Replay bufferReplay Buffer[90, 89]自从被提出后就成了Single-Agent强化学习的常规操作，特别是DQN一炮走红之后[72] 。

不过，Replay Buffer有着很强的理论假设，用原作者的话说是——The environment should not changeover time because this makes pastexperiences irrelevantor even harmful.（环境不应随时间而改变，因为这会使过去的experience replay变得无关紧要甚至有害）Replay buffer假设环境是stationary的，如果当前的环境信息不同于过去的环境信息，那么就无法从过去环境的replay中学到有价值的经验。

（画外音：大人，时代变了……别刻舟求剑了）在multi-agent场景下，每个agent都可以把其他的agent当作环境的一部分。

因为其他的agent不断地学习进化，所以agent所处的环境也是在不断变换的，也就是所谓的non-stationary。

因为multi-agent场景不符合replay buffer的理论假设，所以有的人就直接放弃治疗了——例如2016年发表的大名鼎鼎的RIAL和DIAL中就没有使用replay buffer。

Wireless Sensor Network, 2012, 4, 162-166doi:10.4236/wsn.2012.46023 Published Online June 2012 (/journal/wsn)MNMU-RA: Most Nearest Most Used Routing Algorithm for Greening the Wireless Sensor NetworksHafiz Bilal Khalil, Syed Jawad Hussain ZaidiSchool of Electrical Engineering & Computer Sciences, National University of Sciences and Technology, Islamabad, PakistanEmail: {10mseetkhalil, 10mseejzaidi}@.pkReceived February 22, 2012; revised March 22, 2012; accepted April 10, 2012ABSTRACTWireless sensors are widely deployed in military and other organizations that significantly depend upon the sensed in-formation in any emergency situation. One of the main designs issues of the wireless sensor network (WSN) is the con-servation of energy which is directly proportional to the life of the networks. We propose most nearest most used rout-ing algorithm (MNMU-RA) for ad-hoc WSNs which vitally plays an important role in energy conservation. We find the best location of MNMU node for energy harvesting by apply our algorithm. Our method involves the least number of nodes in transmission of data and set large number of nodes to sleep in idle mode. Based on simulation result we shows the significant improvement in energy saving and enhance the life of the network.Keywords: Energy Efficiency; Wireless Sensor Networks; Routing1. IntroductionThe growth in wireless sensor networks and its applica- tions dramatically increased in last decade. Wireless sen- sor nodes are widely used in military surveillance, intel- ligence and targeting in war operations. Energy available at each sensor for sensing and communications is limited because of the cost constraints and smaller size, which affects the sensor application and network lifetime. The purpose of green networking is to overcome the carbon foot print, reduce the energy consumption and energy losses. Energy efficiency is an important issue to enhance the life time of the network. To achieve the green net- working every component of the network is integrated with energy efficient protocols, e.g., energy-aware rout- ing on network layer, energy-saving mode on MAC layer, etc. One of the most important components of the sensor node is the power source. In sensor networks generally there are three modes of power consumption: sensing, data processing, and communication. Compared to sensing and data processing, much more energy is required for data communication in a typical sensor node [1]. These are also categorized as sleep (idle) and wakeup (trans-mission) mode.In ad-hoc WSNs (Wireless Sensor Networks) always the nodes are cooperative, they sense and transmit their own data and also act as router to route the sensed infor- mation of other nodes towards the data center or gateway node which is connected to the internet. Most of the nodes consumed their power resource while transmitting the data of neighboring nodes. The scope of this paper is to minimize the power consumption in transmitting or routing process and set large number of nodes into sleep mode. The remaining sections of this paper organized as follows. Section 2 explains related work and current en-ergy efficient techniques for sensor networks. Section 3 introduces some problems and research issues in current work. Section 4 describes overview of network model, our proposed algorithm and proposed solution respec-tively. In Section 5 experiment, Results and comparisons are given.2. Related WorkEnergy efficiency is already achieved by many appro- aches. These approaches include energy aware protocol development and hardware optimizations, such as sleep- ing schedules to keep electronics inactive most of the time, dynamic optimization of voltage, and clock rate. In[2] Smart Dust motes are designed that are not more thana few cubic millimeters. They can float in the air, keep sensing and transmitting for hours or days. In [3] authors described the µAMPS wireless sensor node, it is hard- ware based solution in which they simultaneously con- sider the features of the microprocessors and transceivers to reduce the power consumption of the each wireless sensor node in network. Routing algorithms also play an important role to reduce the energy consumption during the routing of data. A lot of work is done in MAC layer and Mac protocols;MAC protocol for wireless sensorH. B. KHALIL, S. J. H. ZAIDI163networks is not like the traditional wireless MACs such as IEEE 802.11. One of the most important goals is en-ergy conservation, fairness and latency is less important [4].SMAC/AL (Sensor MAC with Adaptive Listening) is a famous MAC protocol for WSNs proposed by Ye et al. [5,6]. Main purpose of SMAC/AL is to reduce energy consumption. But in SMAC/AL without considering the distance among the nodes, all nodes unnecessarily con- sume the energy by transmitting information with con- stant power level. An energy efficient MAC protocol with adaptive transmit power scheme named ATPM (Adap- tive Transmit Power MAC) is proposed in [7]. By meas- uring the received power ATPM can calculate the dis- tance between the sender and the receiver, and then adap- tively choose the suitable transmit power level according to the propagation model and distance. The ATMP can not only conserve the energy source, but also decrease the collision probability. A Novel Clustering Algorithm for Energy Efficiency in Wireless Sensor Networks (AN-CAEE) has been proposed [8]. It minimizes energy utili-zation during data transmission and energy consumptions are distributed uniformly among all nodes. Each cluster contains cluster head, each node send its data to cluster head with single hop transmission. And cluster transmits the combined data to the base station with multi hope transmission. This approach reduces energy consumption of nodes within the cluster.3. Problem StatementSensor nodes which are one hope away or closest to the gateway node always consume their power more quickly than others because they have to transmit the data of other nodes in addition to transmission of their own sensed information. In [9] a solution was proposed for such type of scenario by implementing the multiple base stations and periodically changing their positions. But the prob- lem is that if every time the most far away sensor trans- mits its data then major part of overall network energy will be consumed. Another solution for prolong the sen- sor network lifetime is to divide sensors nodes into dis- joint sets, such that all the targets completely covered by every set [3]. Authors consider that within an active sen- sor’s operational range a target is covered. These disjoint sets are activated in round robin fashion, such that at a time only one set is active. Sensors are into the active state in an active set and all other sensors are in a low- energy sleep state. According to this method almost half of the sensor remains active and remaining half goes to sleep mode which reduce energy down to 50%. To make it more efficient and conserve the larger amount of en- ergy we proposed an algorithm named as MNMU-RA (Most nearest most used routing algorithm). That algo- rithm finds the efficient placement of active sensor nodes and set other nodes into sleep mode. An issue is also re- solved by our algorithm, reducing the number of multiple base stations by finding out the best location for the base station without changing its location periodically.4. Synopsis of Our Network ModelIn this paper we deal with the issue of energy efficiency in wireless sensor networks for surveillance of a set of targets with known locality. Scenario of the network is chosen for armed forces purposes like surveillance of the boarder, battle fields and no go areas to acquire the in- formation about enemies and their locations without tak- ing the risk for human personal. We consider that a large number of sensors are distributed randomly in close prox- imity for monitoring and send the monitored information to a gateway node. All nodes are static and makes ad-hoc wireless sensor network. Every sensor nodes must moni-tor the area all the time in its operational range and each sensor has fixed transmission range. In network model we assume that each sensor has unique pre configured Id and Global/proactive routing algorithms are used. Main advantage of proactive algorithm is not route latency but drawback is the high maintenance overhead when many of the routes are never used.Proactive routing is appro-priate for networks with: Small size, low mobility and high communication rates. We proposed an algorithm called as most nearest most used routing algorithm for this purpose. By using MNMU-RA we can find the per-fect location of node for energy harvesting which also reduce the overall energy consumption and cost.4.1. Most Nearest Most Used Routing Algorithm Run shortest path routing algorithm or link state routing to find the shortest path for each node in the wireless sensor network. Calculate all the possible shortest paths for each node. Then find the MNMU node (Figure 1).∙ A node which is most nearest to the gateway node.∙Select a node which is used in maximum number of shortest paths.Figure 1. Location of selected MNMU node.H. B. KHALIL, S. J. H. ZAIDI 164In above network model we assumed that sensed in- formation is equally probable for all the nodes. Then we calculate the shortest path for the nodes A, B and C. Then we find out the nodes which are most nearest to the gate way node. In above network model there are only two nodes X and Y which are closer to the gateway node. Then for selection we give the preference to the node which is most used in shortest paths. In above model Y is node which is most used in all shortest paths. If nodes A, B and C transmit their data the entire time node Y will be included in their path. Then every node keeps its routes information towards the node Y for future communica- tions. Flow chart of our algorithm is given in Figure 2. 4.2. Proposed SolutionWe used our algorithm to find most nearest most used node in a network, that node should be active all the time while other sensors remain in sleep mode and keep sens- ing. As we use proactive routing so each sensor knows its path towards the MNMU node. If a node has to send its information before sending it will wake up the nodes along his route. When MNMU nodes receive the infor- mation it will forward the data to the gateway and sets all the nodes into sleep mode. The critical issue in this solu- tion is that if a node (MNMU node) remains active all the time then its energy source will be empty soon. We re- solve this issue by using the energy harvesting concept at MNMU node [1]. We can also use secondary batteryFigure 2. MNMU routing algorithm flow chart. which is rechargeable and coupled with photovoltaic cell[10]. If all the nodes can generate energy from light, vi-bration, heat etc [11,12] it will increase the system cost.We don’t need to replace all the nodes with secondary sources. By replacing only one node (MNMU node) re-solves the issue and slightly increases the cost of theoverall system. But effectively prolong the life time ofsensor network. A solution given by Gandham et al. [9]can be more energy efficient if we implement our pro-posed algorithm with every new location of mobile basestation. Split the network in equal parts and periodicallychange the position of base station in each part. Basestation can be easily implemented at the place of MNMUnode in each part of the network instead of replacing itoutside the network. MNMU node will reduce the multihop and number of transmission which directly reducethe energy consumption.5. ExperimentWe done the experiment by implementing our proposedalgorithm in a network and calculate the amount of en-ergy utilization using MATLAB. Then implement theconcept of disjoint set and analyze the values at same network. For simulation 20 nodes containing one gate-way node are distributed randomly in 30 meter squarearea. We consider the features of MICA2 mote platform.It is third generation mote specifically built for WSNs [4].MICA2 have selectable transmission power range whichoffers adjustable communication ranges, selected trans-mission range for each node is 10 meters. The packetlength is fixed at 200 bits. MICA2 usually operated with3 V battery and other features mentioned in Table 1.We divided our analysis in three parts; first we calcu-late the power consumption using disjoint sets methods[3], then we apply our algorithm and calculate & com-pare power consumption. Same network and topologytaken in which each node remains active all the time andno energy saving protocol and technique is implemented.Energy calculated during the 20 rounds, all nodes areactive in first five rounds in which they sense and trans-mit the data. After ten rounds there is no activity andnodes go to sleeping mode according to implemented Table 1. Features of MICA2 motes platform [12,13].Operation/Features UnitListening 8mA Receiving 10mA Transmission 17mA Sleep 19µA Radio Frequency 900 MHzCPU 8 bit Atmel at 8 MHzBandwidth 40KbpsH. B. KHALIL, S. J. H. ZAIDI165methodology. Calculated results are given in Figures3 and 4.Simulation ResultsFigure 3 shows the result comparison of energy con- sumption in different modes; sensing, Transmission and sleeping of network. In Figure 3(a) set of all the active nodes shown by blue line are transmitting the data with- out applying any energy saving protocol. During the transmission if all nodes are active they will keep trans- mitting the information to each other and maximum amount of energy is consumed. In disjoint system only active set take part in transmission and inactive nodesFigure 3. Power consumptions in different modes. (a) Trans- mission mode; (b) Power consume by sleeping nodes; (c)Power consume by active nodes in sleep mode. Figure 4. Result and comparison of energy consumption in different modes.remain inactive during the transmission of active set. Our proposed algorithm gives lowest amount of energy con- sumption because only the MNMU node and less number of nodes take part in transmission. Energy consumed by inactive nodes in sleeping modes is shown in Figure 3(b). Energy consumption of sleeping nodes is in µwatts. Ac- cording to our algorithm 19 nodes set to sleep mode and only one MNMU node is active. While Figure 3(c) shows the separately calculated energy consumption by active nodes when there is no activity and network is in idle mode. Similarly in sleeping mode only MNMU node remains active and rest of the network sets to sleep mode. Figure 4 shows the result of energy consumption of entire network in different rounds. In first 5 rounds we assume that there is no sensed information to send; all the nodes are active in listening mode and keep sensing. In 5 to 10 rounds nodes are transmitting their sensed in- formation to the gateway. After round 10 there is no ac- tivity and nodes set to sleep mode in sleep mode only energy consumed by active nodes are calculated and en- ergy consumed by sleeping nodes which is in µwatts is neglected. Our algorithm gives the minimum energy con- sumption during the transmission in which fewer num- bers of nodes take part in routing and also in sleep mode by keeping only MNMU node active.6. ConclusionWe presented the most nearest most used routing algo- rithm to reduce the energy utilization in wireless sensor networks. Using this algorithm we find the best location of energy harvested node in a network. Our algorithm involves least number of nodes during transmission and keeps one node active in sleep mode. That significantly reduces the energy consumption during the transmissionH. B. KHALIL, S. J. H. ZAIDI 166and sleep mode when there is no activity. An open re- search issue is the heterogeneity of energy resources of the nodes that must be resolved after practical imple- mentation in any network. In our solution there is uneven energy consumption due to the topology of the network and nature of data flow. But that uneven energy con- sumption is helpful to reduce the energy consumption of entire network7. Future DirectionDesired goal in wireless networks is energy efficiency to maximize the network life. Our algorithm can be used to find the location of cluster header quickly in novel clus- tering algorithm for energy efficiency in wireless sensor networks [8]. Further we can implement coding tech- niques to reduce the number of transmissions at MNMU node. Energy consumes per bit or per packet transmis- sion can be reduce. Number of packets can be transmit- ted as a single packet by applying x-or Operations which reduces the energy but may cause of slighter delay. To apply this technique sensor nodes must be smarter and have ability to do this quickly.REFERENCES[1]I. F. Akyildiz, T. Melodia and K. Chowdhury, “A Surveyon Wireless Multimedia Sensor Networks,” ComputerNetworks, Vol. 51, No. 4, 2007, pp. 921-960.doi:10.1016/net.2006.10.002[2]J. M. Kahn, R. H. Katz and K. S. J. Pister, “EmergingChallenges: Mobile Networking for Smart Dust,” Inter-national Journal of Communication Networks, Vol. 2, No.3, 2000, pp. 188-196.[3]M. Cardei and D. Z. Du, “Improving Wireless SensorNetwork Lifetime through Power Aware Organization,”Wireless Networks, Vol. 11, No. 3, 2005, pp. 333-340.doi:10.1007/s11276-005-6615-6[4]Q. Hu and Z. Z. Tang, “An Adaptive Transmit PowerScheme for Wireless Sensor Networks,” 3rd IEEE Inter-national Conference on Ubi-Media Computing, Jinhua, 5-7 July 2010, pp. 12-16.[5]W. Ye, J. Heidemann and D. Estrin, “An Energy-EfficientMAC Protocol for Wireless Sensor Networks,” Proceed- ings of the IEEE INFOCOM, New York, 23-27 June 2002, pp. 1567-1576.[6]W. Ye, J. Heidemann and D. Estrin, “Medium AccessControl with Coordinated Adaptive Sleeping for Wireless Sensor Networks,” IEEE/ACM Transactions on Network- ing, Vol. 12, No. 3, 2004, pp. 493-506.doi:10.1109/TNET.2004.828953[7]Q. Hu and Z. Tang, “ATPM: An Energy Efficient MACProtocol with Adaptive Transmit Power Scheme for Wire- less Sensor Networks,” Journal of Multimedia, Vol. 6, No.2, 2011, pp. 122-128. doi:10.4304/jmm.6.2.122-128[8] A. P. Abidoye and N. A. Azeez, “ANCAEE: A Novel Clus-tering Algorithm for Energy Efficiency in Wireless Sen- sor Networks,” Journal of Wireless Sensor Networks, Vol.3, No. 9, 2011, pp. 307-312. doi:10.4236/wsn.2011.39032 [9]S. R. Gandham, M. Dawande, R. Prakash and S. Venkate-san, “Energy Efficient Schemes for Wireless Sensor Net- works with Multiple Mobile Base Stations,” Global Tele- communications Conference, San Francisco, 1-5 Decem- ber 2003, pp. 377-381.[10]M. A. M. Vieira, C. N. Coelho, D. C. Silva and J. M. Mata,“Survey on Wireless Sensor Network Devices,” Proceed- ings of IEEE International Conference on Emerging Tec- hnologies and Factory Automation (ETFA’03), Lisbon, 16-19 September 2003, pp. 537-544.[11]J. Paradiso and T. Starner, “Energy Scavenging for Mo-bile and Wireless Electronics,” Pervasive Computing, Vol.4, No. 1, 2005, pp. 18-27. doi:10.1109/MPRV.2005.9 [12]V. Gungor and G. Hancke, “Industrial Wireless SensorNetworks: Challenges, Design Principles, and Technical Approaches,” IEEE Transactions on Industrial Electron- ics, Vol. 56, No. 10, 2009, pp. 4258-4265.doi:10.1109/TIE.2009.2015754[13]CrossBow, Mica2 Data Sheet./Products/Product_pdf_files/MICA%20data%20sheet.pdf。

机器学习与人工智能领域中常用的英语词汇1.General Concepts (基础概念)•Artificial Intelligence (AI) - 人工智能1)Artificial Intelligence (AI) - 人工智能2)Machine Learning (ML) - 机器学习3)Deep Learning (DL) - 深度学习4)Neural Network - 神经网络5)Natural Language Processing (NLP) - 自然语言处理6)Computer Vision - 计算机视觉7)Robotics - 机器人技术8)Speech Recognition - 语音识别9)Expert Systems - 专家系统10)Knowledge Representation - 知识表示11)Pattern Recognition - 模式识别12)Cognitive Computing - 认知计算13)Autonomous Systems - 自主系统14)Human-Machine Interaction - 人机交互15)Intelligent Agents - 智能代理16)Machine Translation - 机器翻译17)Swarm Intelligence - 群体智能18)Genetic Algorithms - 遗传算法19)Fuzzy Logic - 模糊逻辑20)Reinforcement Learning - 强化学习•Machine Learning (ML) - 机器学习1)Machine Learning (ML) - 机器学习2)Artificial Neural Network - 人工神经网络3)Deep Learning - 深度学习4)Supervised Learning - 有监督学习5)Unsupervised Learning - 无监督学习6)Reinforcement Learning - 强化学习7)Semi-Supervised Learning - 半监督学习8)Training Data - 训练数据9)Test Data - 测试数据10)Validation Data - 验证数据11)Feature - 特征12)Label - 标签13)Model - 模型14)Algorithm - 算法15)Regression - 回归16)Classification - 分类17)Clustering - 聚类18)Dimensionality Reduction - 降维19)Overfitting - 过拟合20)Underfitting - 欠拟合•Deep Learning (DL) - 深度学习1)Deep Learning - 深度学习2)Neural Network - 神经网络3)Artificial Neural Network (ANN) - 人工神经网络4)Convolutional Neural Network (CNN) - 卷积神经网络5)Recurrent Neural Network (RNN) - 循环神经网络6)Long Short-Term Memory (LSTM) - 长短期记忆网络7)Gated Recurrent Unit (GRU) - 门控循环单元8)Autoencoder - 自编码器9)Generative Adversarial Network (GAN) - 生成对抗网络10)Transfer Learning - 迁移学习11)Pre-trained Model - 预训练模型12)Fine-tuning - 微调13)Feature Extraction - 特征提取14)Activation Function - 激活函数15)Loss Function - 损失函数16)Gradient Descent - 梯度下降17)Backpropagation - 反向传播18)Epoch - 训练周期19)Batch Size - 批量大小20)Dropout - 丢弃法•Neural Network - 神经网络1)Neural Network - 神经网络2)Artificial Neural Network (ANN) - 人工神经网络3)Deep Neural Network (DNN) - 深度神经网络4)Convolutional Neural Network (CNN) - 卷积神经网络5)Recurrent Neural Network (RNN) - 循环神经网络6)Long Short-Term Memory (LSTM) - 长短期记忆网络7)Gated Recurrent Unit (GRU) - 门控循环单元8)Feedforward Neural Network - 前馈神经网络9)Multi-layer Perceptron (MLP) - 多层感知器10)Radial Basis Function Network (RBFN) - 径向基函数网络11)Hopfield Network - 霍普菲尔德网络12)Boltzmann Machine - 玻尔兹曼机13)Autoencoder - 自编码器14)Spiking Neural Network (SNN) - 脉冲神经网络15)Self-organizing Map (SOM) - 自组织映射16)Restricted Boltzmann Machine (RBM) - 受限玻尔兹曼机17)Hebbian Learning - 海比安学习18)Competitive Learning - 竞争学习19)Neuroevolutionary - 神经进化20)Neuron - 神经元•Algorithm - 算法1)Algorithm - 算法2)Supervised Learning Algorithm - 有监督学习算法3)Unsupervised Learning Algorithm - 无监督学习算法4)Reinforcement Learning Algorithm - 强化学习算法5)Classification Algorithm - 分类算法6)Regression Algorithm - 回归算法7)Clustering Algorithm - 聚类算法8)Dimensionality Reduction Algorithm - 降维算法9)Decision Tree Algorithm - 决策树算法10)Random Forest Algorithm - 随机森林算法11)Support Vector Machine (SVM) Algorithm - 支持向量机算法12)K-Nearest Neighbors (KNN) Algorithm - K近邻算法13)Naive Bayes Algorithm - 朴素贝叶斯算法14)Gradient Descent Algorithm - 梯度下降算法15)Genetic Algorithm - 遗传算法16)Neural Network Algorithm - 神经网络算法17)Deep Learning Algorithm - 深度学习算法18)Ensemble Learning Algorithm - 集成学习算法19)Reinforcement Learning Algorithm - 强化学习算法20)Metaheuristic Algorithm - 元启发式算法•Model - 模型1)Model - 模型2)Machine Learning Model - 机器学习模型3)Artificial Intelligence Model - 人工智能模型4)Predictive Model - 预测模型5)Classification Model - 分类模型6)Regression Model - 回归模型7)Generative Model - 生成模型8)Discriminative Model - 判别模型9)Probabilistic Model - 概率模型10)Statistical Model - 统计模型11)Neural Network Model - 神经网络模型12)Deep Learning Model - 深度学习模型13)Ensemble Model - 集成模型14)Reinforcement Learning Model - 强化学习模型15)Support Vector Machine (SVM) Model - 支持向量机模型16)Decision Tree Model - 决策树模型17)Random Forest Model - 随机森林模型18)Naive Bayes Model - 朴素贝叶斯模型19)Autoencoder Model - 自编码器模型20)Convolutional Neural Network (CNN) Model - 卷积神经网络模型•Dataset - 数据集1)Dataset - 数据集2)Training Dataset - 训练数据集3)Test Dataset - 测试数据集4)Validation Dataset - 验证数据集5)Balanced Dataset - 平衡数据集6)Imbalanced Dataset - 不平衡数据集7)Synthetic Dataset - 合成数据集8)Benchmark Dataset - 基准数据集9)Open Dataset - 开放数据集10)Labeled Dataset - 标记数据集11)Unlabeled Dataset - 未标记数据集12)Semi-Supervised Dataset - 半监督数据集13)Multiclass Dataset - 多分类数据集14)Feature Set - 特征集15)Data Augmentation - 数据增强16)Data Preprocessing - 数据预处理17)Missing Data - 缺失数据18)Outlier Detection - 异常值检测19)Data Imputation - 数据插补20)Metadata - 元数据•Training - 训练1)Training - 训练2)Training Data - 训练数据3)Training Phase - 训练阶段4)Training Set - 训练集5)Training Examples - 训练样本6)Training Instance - 训练实例7)Training Algorithm - 训练算法8)Training Model - 训练模型9)Training Process - 训练过程10)Training Loss - 训练损失11)Training Epoch - 训练周期12)Training Batch - 训练批次13)Online Training - 在线训练14)Offline Training - 离线训练15)Continuous Training - 连续训练16)Transfer Learning - 迁移学习17)Fine-Tuning - 微调18)Curriculum Learning - 课程学习19)Self-Supervised Learning - 自监督学习20)Active Learning - 主动学习•Testing - 测试1)Testing - 测试2)Test Data - 测试数据3)Test Set - 测试集4)Test Examples - 测试样本5)Test Instance - 测试实例6)Test Phase - 测试阶段7)Test Accuracy - 测试准确率8)Test Loss - 测试损失9)Test Error - 测试错误10)Test Metrics - 测试指标11)Test Suite - 测试套件12)Test Case - 测试用例13)Test Coverage - 测试覆盖率14)Cross-Validation - 交叉验证15)Holdout Validation - 留出验证16)K-Fold Cross-Validation - K折交叉验证17)Stratified Cross-Validation - 分层交叉验证18)Test Driven Development (TDD) - 测试驱动开发19)A/B Testing - A/B 测试20)Model Evaluation - 模型评估•Validation - 验证1)Validation - 验证2)Validation Data - 验证数据3)Validation Set - 验证集4)Validation Examples - 验证样本5)Validation Instance - 验证实例6)Validation Phase - 验证阶段7)Validation Accuracy - 验证准确率8)Validation Loss - 验证损失9)Validation Error - 验证错误10)Validation Metrics - 验证指标11)Cross-Validation - 交叉验证12)Holdout Validation - 留出验证13)K-Fold Cross-Validation - K折交叉验证14)Stratified Cross-Validation - 分层交叉验证15)Leave-One-Out Cross-Validation - 留一法交叉验证16)Validation Curve - 验证曲线17)Hyperparameter Validation - 超参数验证18)Model Validation - 模型验证19)Early Stopping - 提前停止20)Validation Strategy - 验证策略•Supervised Learning - 有监督学习1)Supervised Learning - 有监督学习2)Label - 标签3)Feature - 特征4)Target - 目标5)Training Labels - 训练标签6)Training Features - 训练特征7)Training Targets - 训练目标8)Training Examples - 训练样本9)Training Instance - 训练实例10)Regression - 回归11)Classification - 分类12)Predictor - 预测器13)Regression Model - 回归模型14)Classifier - 分类器15)Decision Tree - 决策树16)Support Vector Machine (SVM) - 支持向量机17)Neural Network - 神经网络18)Feature Engineering - 特征工程19)Model Evaluation - 模型评估20)Overfitting - 过拟合21)Underfitting - 欠拟合22)Bias-Variance Tradeoff - 偏差-方差权衡•Unsupervised Learning - 无监督学习1)Unsupervised Learning - 无监督学习2)Clustering - 聚类3)Dimensionality Reduction - 降维4)Anomaly Detection - 异常检测5)Association Rule Learning - 关联规则学习6)Feature Extraction - 特征提取7)Feature Selection - 特征选择8)K-Means - K均值9)Hierarchical Clustering - 层次聚类10)Density-Based Clustering - 基于密度的聚类11)Principal Component Analysis (PCA) - 主成分分析12)Independent Component Analysis (ICA) - 独立成分分析13)T-distributed Stochastic Neighbor Embedding (t-SNE) - t分布随机邻居嵌入14)Gaussian Mixture Model (GMM) - 高斯混合模型15)Self-Organizing Maps (SOM) - 自组织映射16)Autoencoder - 自动编码器17)Latent Variable - 潜变量18)Data Preprocessing - 数据预处理19)Outlier Detection - 异常值检测20)Clustering Algorithm - 聚类算法•Reinforcement Learning - 强化学习1)Reinforcement Learning - 强化学习2)Agent - 代理3)Environment - 环境4)State - 状态5)Action - 动作6)Reward - 奖励7)Policy - 策略8)Value Function - 值函数9)Q-Learning - Q学习10)Deep Q-Network (DQN) - 深度Q网络11)Policy Gradient - 策略梯度12)Actor-Critic - 演员-评论家13)Exploration - 探索14)Exploitation - 开发15)Temporal Difference (TD) - 时间差分16)Markov Decision Process (MDP) - 马尔可夫决策过程17)State-Action-Reward-State-Action (SARSA) - 状态-动作-奖励-状态-动作18)Policy Iteration - 策略迭代19)Value Iteration - 值迭代20)Monte Carlo Methods - 蒙特卡洛方法•Semi-Supervised Learning - 半监督学习1)Semi-Supervised Learning - 半监督学习2)Labeled Data - 有标签数据3)Unlabeled Data - 无标签数据4)Label Propagation - 标签传播5)Self-Training - 自训练6)Co-Training - 协同训练7)Transudative Learning - 传导学习8)Inductive Learning - 归纳学习9)Manifold Regularization - 流形正则化10)Graph-based Methods - 基于图的方法11)Cluster Assumption - 聚类假设12)Low-Density Separation - 低密度分离13)Semi-Supervised Support Vector Machines (S3VM) - 半监督支持向量机14)Expectation-Maximization (EM) - 期望最大化15)Co-EM - 协同期望最大化16)Entropy-Regularized EM - 熵正则化EM17)Mean Teacher - 平均教师18)Virtual Adversarial Training - 虚拟对抗训练19)Tri-training - 三重训练20)Mix Match - 混合匹配•Feature - 特征1)Feature - 特征2)Feature Engineering - 特征工程3)Feature Extraction - 特征提取4)Feature Selection - 特征选择5)Input Features - 输入特征6)Output Features - 输出特征7)Feature Vector - 特征向量8)Feature Space - 特征空间9)Feature Representation - 特征表示10)Feature Transformation - 特征转换11)Feature Importance - 特征重要性12)Feature Scaling - 特征缩放13)Feature Normalization - 特征归一化14)Feature Encoding - 特征编码15)Feature Fusion - 特征融合16)Feature Dimensionality Reduction - 特征维度减少17)Continuous Feature - 连续特征18)Categorical Feature - 分类特征19)Nominal Feature - 名义特征20)Ordinal Feature - 有序特征•Label - 标签1)Label - 标签2)Labeling - 标注3)Ground Truth - 地面真值4)Class Label - 类别标签5)Target Variable - 目标变量6)Labeling Scheme - 标注方案7)Multi-class Labeling - 多类别标注8)Binary Labeling - 二分类标注9)Label Noise - 标签噪声10)Labeling Error - 标注错误11)Label Propagation - 标签传播12)Unlabeled Data - 无标签数据13)Labeled Data - 有标签数据14)Semi-supervised Learning - 半监督学习15)Active Learning - 主动学习16)Weakly Supervised Learning - 弱监督学习17)Noisy Label Learning - 噪声标签学习18)Self-training - 自训练19)Crowdsourcing Labeling - 众包标注20)Label Smoothing - 标签平滑化•Prediction - 预测1)Prediction - 预测2)Forecasting - 预测3)Regression - 回归4)Classification - 分类5)Time Series Prediction - 时间序列预测6)Forecast Accuracy - 预测准确性7)Predictive Modeling - 预测建模8)Predictive Analytics - 预测分析9)Forecasting Method - 预测方法10)Predictive Performance - 预测性能11)Predictive Power - 预测能力12)Prediction Error - 预测误差13)Prediction Interval - 预测区间14)Prediction Model - 预测模型15)Predictive Uncertainty - 预测不确定性16)Forecast Horizon - 预测时间跨度17)Predictive Maintenance - 预测性维护18)Predictive Policing - 预测式警务19)Predictive Healthcare - 预测性医疗20)Predictive Maintenance - 预测性维护•Classification - 分类1)Classification - 分类2)Classifier - 分类器3)Class - 类别4)Classify - 对数据进行分类5)Class Label - 类别标签6)Binary Classification - 二元分类7)Multiclass Classification - 多类分类8)Class Probability - 类别概率9)Decision Boundary - 决策边界10)Decision Tree - 决策树11)Support Vector Machine (SVM) - 支持向量机12)K-Nearest Neighbors (KNN) - K最近邻算法13)Naive Bayes - 朴素贝叶斯14)Logistic Regression - 逻辑回归15)Random Forest - 随机森林16)Neural Network - 神经网络17)SoftMax Function - SoftMax函数18)One-vs-All (One-vs-Rest) - 一对多(一对剩余)19)Ensemble Learning - 集成学习20)Confusion Matrix - 混淆矩阵•Regression - 回归1)Regression Analysis - 回归分析2)Linear Regression - 线性回归3)Multiple Regression - 多元回归4)Polynomial Regression - 多项式回归5)Logistic Regression - 逻辑回归6)Ridge Regression - 岭回归7)Lasso Regression - Lasso回归8)Elastic Net Regression - 弹性网络回归9)Regression Coefficients - 回归系数10)Residuals - 残差11)Ordinary Least Squares (OLS) - 普通最小二乘法12)Ridge Regression Coefficient - 岭回归系数13)Lasso Regression Coefficient - Lasso回归系数14)Elastic Net Regression Coefficient - 弹性网络回归系数15)Regression Line - 回归线16)Prediction Error - 预测误差17)Regression Model - 回归模型18)Nonlinear Regression - 非线性回归19)Generalized Linear Models (GLM) - 广义线性模型20)Coefficient of Determination (R-squared) - 决定系数21)F-test - F检验22)Homoscedasticity - 同方差性23)Heteroscedasticity - 异方差性24)Autocorrelation - 自相关25)Multicollinearity - 多重共线性26)Outliers - 异常值27)Cross-validation - 交叉验证28)Feature Selection - 特征选择29)Feature Engineering - 特征工程30)Regularization - 正则化2.Neural Networks and Deep Learning (神经网络与深度学习)•Convolutional Neural Network (CNN) - 卷积神经网络1)Convolutional Neural Network (CNN) - 卷积神经网络2)Convolution Layer - 卷积层3)Feature Map - 特征图4)Convolution Operation - 卷积操作5)Stride - 步幅6)Padding - 填充7)Pooling Layer - 池化层8)Max Pooling - 最大池化9)Average Pooling - 平均池化10)Fully Connected Layer - 全连接层11)Activation Function - 激活函数12)Rectified Linear Unit (ReLU) - 线性修正单元13)Dropout - 随机失活14)Batch Normalization - 批量归一化15)Transfer Learning - 迁移学习16)Fine-Tuning - 微调17)Image Classification - 图像分类18)Object Detection - 物体检测19)Semantic Segmentation - 语义分割20)Instance Segmentation - 实例分割21)Generative Adversarial Network (GAN) - 生成对抗网络22)Image Generation - 图像生成23)Style Transfer - 风格迁移24)Convolutional Autoencoder - 卷积自编码器25)Recurrent Neural Network (RNN) - 循环神经网络•Recurrent Neural Network (RNN) - 循环神经网络1)Recurrent Neural Network (RNN) - 循环神经网络2)Long Short-Term Memory (LSTM) - 长短期记忆网络3)Gated Recurrent Unit (GRU) - 门控循环单元4)Sequence Modeling - 序列建模5)Time Series Prediction - 时间序列预测6)Natural Language Processing (NLP) - 自然语言处理7)Text Generation - 文本生成8)Sentiment Analysis - 情感分析9)Named Entity Recognition (NER) - 命名实体识别10)Part-of-Speech Tagging (POS Tagging) - 词性标注11)Sequence-to-Sequence (Seq2Seq) - 序列到序列12)Attention Mechanism - 注意力机制13)Encoder-Decoder Architecture - 编码器-解码器架构14)Bidirectional RNN - 双向循环神经网络15)Teacher Forcing - 强制教师法16)Backpropagation Through Time (BPTT) - 通过时间的反向传播17)Vanishing Gradient Problem - 梯度消失问题18)Exploding Gradient Problem - 梯度爆炸问题19)Language Modeling - 语言建模20)Speech Recognition - 语音识别•Long Short-Term Memory (LSTM) - 长短期记忆网络1)Long Short-Term Memory (LSTM) - 长短期记忆网络2)Cell State - 细胞状态3)Hidden State - 隐藏状态4)Forget Gate - 遗忘门5)Input Gate - 输入门6)Output Gate - 输出门7)Peephole Connections - 窥视孔连接8)Gated Recurrent Unit (GRU) - 门控循环单元9)Vanishing Gradient Problem - 梯度消失问题10)Exploding Gradient Problem - 梯度爆炸问题11)Sequence Modeling - 序列建模12)Time Series Prediction - 时间序列预测13)Natural Language Processing (NLP) - 自然语言处理14)Text Generation - 文本生成15)Sentiment Analysis - 情感分析16)Named Entity Recognition (NER) - 命名实体识别17)Part-of-Speech Tagging (POS Tagging) - 词性标注18)Attention Mechanism - 注意力机制19)Encoder-Decoder Architecture - 编码器-解码器架构20)Bidirectional LSTM - 双向长短期记忆网络•Attention Mechanism - 注意力机制1)Attention Mechanism - 注意力机制2)Self-Attention - 自注意力3)Multi-Head Attention - 多头注意力4)Transformer - 变换器5)Query - 查询6)Key - 键7)Value - 值8)Query-Value Attention - 查询-值注意力9)Dot-Product Attention - 点积注意力10)Scaled Dot-Product Attention - 缩放点积注意力11)Additive Attention - 加性注意力12)Context Vector - 上下文向量13)Attention Score - 注意力分数14)SoftMax Function - SoftMax函数15)Attention Weight - 注意力权重16)Global Attention - 全局注意力17)Local Attention - 局部注意力18)Positional Encoding - 位置编码19)Encoder-Decoder Attention - 编码器-解码器注意力20)Cross-Modal Attention - 跨模态注意力•Generative Adversarial Network (GAN) - 生成对抗网络1)Generative Adversarial Network (GAN) - 生成对抗网络2)Generator - 生成器3)Discriminator - 判别器4)Adversarial Training - 对抗训练5)Minimax Game - 极小极大博弈6)Nash Equilibrium - 纳什均衡7)Mode Collapse - 模式崩溃8)Training Stability - 训练稳定性9)Loss Function - 损失函数10)Discriminative Loss - 判别损失11)Generative Loss - 生成损失12)Wasserstein GAN (WGAN) - Wasserstein GAN（WGAN）13)Deep Convolutional GAN (DCGAN) - 深度卷积生成对抗网络（DCGAN）14)Conditional GAN (c GAN) - 条件生成对抗网络（c GAN）15)Style GAN - 风格生成对抗网络16)Cycle GAN - 循环生成对抗网络17)Progressive Growing GAN (PGGAN) - 渐进式增长生成对抗网络（PGGAN）18)Self-Attention GAN (SAGAN) - 自注意力生成对抗网络（SAGAN）19)Big GAN - 大规模生成对抗网络20)Adversarial Examples - 对抗样本•Encoder-Decoder - 编码器-解码器1)Encoder-Decoder Architecture - 编码器-解码器架构2)Encoder - 编码器3)Decoder - 解码器4)Sequence-to-Sequence Model (Seq2Seq) - 序列到序列模型5)State Vector - 状态向量6)Context Vector - 上下文向量7)Hidden State - 隐藏状态8)Attention Mechanism - 注意力机制9)Teacher Forcing - 强制教师法10)Beam Search - 束搜索11)Recurrent Neural Network (RNN) - 循环神经网络12)Long Short-Term Memory (LSTM) - 长短期记忆网络13)Gated Recurrent Unit (GRU) - 门控循环单元14)Bidirectional Encoder - 双向编码器15)Greedy Decoding - 贪婪解码16)Masking - 遮盖17)Dropout - 随机失活18)Embedding Layer - 嵌入层19)Cross-Entropy Loss - 交叉熵损失20)Tokenization - 令牌化•Transfer Learning - 迁移学习1)Transfer Learning - 迁移学习2)Source Domain - 源领域3)Target Domain - 目标领域4)Fine-Tuning - 微调5)Domain Adaptation - 领域自适应6)Pre-Trained Model - 预训练模型7)Feature Extraction - 特征提取8)Knowledge Transfer - 知识迁移9)Unsupervised Domain Adaptation - 无监督领域自适应10)Semi-Supervised Domain Adaptation - 半监督领域自适应11)Multi-Task Learning - 多任务学习12)Data Augmentation - 数据增强13)Task Transfer - 任务迁移14)Model Agnostic Meta-Learning (MAML) - 与模型无关的元学习（MAML）15)One-Shot Learning - 单样本学习16)Zero-Shot Learning - 零样本学习17)Few-Shot Learning - 少样本学习18)Knowledge Distillation - 知识蒸馏19)Representation Learning - 表征学习20)Adversarial Transfer Learning - 对抗迁移学习•Pre-trained Models - 预训练模型1)Pre-trained Model - 预训练模型2)Transfer Learning - 迁移学习3)Fine-Tuning - 微调4)Knowledge Transfer - 知识迁移5)Domain Adaptation - 领域自适应6)Feature Extraction - 特征提取7)Representation Learning - 表征学习8)Language Model - 语言模型9)Bidirectional Encoder Representations from Transformers (BERT) - 双向编码器结构转换器10)Generative Pre-trained Transformer (GPT) - 生成式预训练转换器11)Transformer-based Models - 基于转换器的模型12)Masked Language Model (MLM) - 掩蔽语言模型13)Cloze Task - 填空任务14)Tokenization - 令牌化15)Word Embeddings - 词嵌入16)Sentence Embeddings - 句子嵌入17)Contextual Embeddings - 上下文嵌入18)Self-Supervised Learning - 自监督学习19)Large-Scale Pre-trained Models - 大规模预训练模型•Loss Function - 损失函数1)Loss Function - 损失函数2)Mean Squared Error (MSE) - 均方误差3)Mean Absolute Error (MAE) - 平均绝对误差4)Cross-Entropy Loss - 交叉熵损失5)Binary Cross-Entropy Loss - 二元交叉熵损失6)Categorical Cross-Entropy Loss - 分类交叉熵损失7)Hinge Loss - 合页损失8)Huber Loss - Huber损失9)Wasserstein Distance - Wasserstein距离10)Triplet Loss - 三元组损失11)Contrastive Loss - 对比损失12)Dice Loss - Dice损失13)Focal Loss - 焦点损失14)GAN Loss - GAN损失15)Adversarial Loss - 对抗损失16)L1 Loss - L1损失17)L2 Loss - L2损失18)Huber Loss - Huber损失19)Quantile Loss - 分位数损失•Activation Function - 激活函数1)Activation Function - 激活函数2)Sigmoid Function - Sigmoid函数3)Hyperbolic Tangent Function (Tanh) - 双曲正切函数4)Rectified Linear Unit (Re LU) - 矩形线性单元5)Parametric Re LU (P Re LU) - 参数化Re LU6)Exponential Linear Unit (ELU) - 指数线性单元7)Swish Function - Swish函数8)Softplus Function - Soft plus函数9)Softmax Function - SoftMax函数10)Hard Tanh Function - 硬双曲正切函数11)Softsign Function - Softsign函数12)GELU (Gaussian Error Linear Unit) - GELU（高斯误差线性单元）13)Mish Function - Mish函数14)CELU (Continuous Exponential Linear Unit) - CELU（连续指数线性单元）15)Bent Identity Function - 弯曲恒等函数16)Gaussian Error Linear Units (GELUs) - 高斯误差线性单元17)Adaptive Piecewise Linear (APL) - 自适应分段线性函数18)Radial Basis Function (RBF) - 径向基函数•Backpropagation - 反向传播1)Backpropagation - 反向传播2)Gradient Descent - 梯度下降3)Partial Derivative - 偏导数4)Chain Rule - 链式法则5)Forward Pass - 前向传播6)Backward Pass - 反向传播7)Computational Graph - 计算图8)Neural Network - 神经网络9)Loss Function - 损失函数10)Gradient Calculation - 梯度计算11)Weight Update - 权重更新12)Activation Function - 激活函数13)Optimizer - 优化器14)Learning Rate - 学习率15)Mini-Batch Gradient Descent - 小批量梯度下降16)Stochastic Gradient Descent (SGD) - 随机梯度下降17)Batch Gradient Descent - 批量梯度下降18)Momentum - 动量19)Adam Optimizer - Adam优化器20)Learning Rate Decay - 学习率衰减•Gradient Descent - 梯度下降1)Gradient Descent - 梯度下降2)Stochastic Gradient Descent (SGD) - 随机梯度下降3)Mini-Batch Gradient Descent - 小批量梯度下降4)Batch Gradient Descent - 批量梯度下降5)Learning Rate - 学习率6)Momentum - 动量7)Adaptive Moment Estimation (Adam) - 自适应矩估计8)RMSprop - 均方根传播9)Learning Rate Schedule - 学习率调度10)Convergence - 收敛11)Divergence - 发散12)Adagrad - 自适应学习速率方法13)Adadelta - 自适应增量学习率方法14)Adamax - 自适应矩估计的扩展版本15)Nadam - Nesterov Accelerated Adaptive Moment Estimation16)Learning Rate Decay - 学习率衰减17)Step Size - 步长18)Conjugate Gradient Descent - 共轭梯度下降19)Line Search - 线搜索20)Newton's Method - 牛顿法•Learning Rate - 学习率1)Learning Rate - 学习率2)Adaptive Learning Rate - 自适应学习率3)Learning Rate Decay - 学习率衰减4)Initial Learning Rate - 初始学习率5)Step Size - 步长6)Momentum - 动量7)Exponential Decay - 指数衰减8)Annealing - 退火9)Cyclical Learning Rate - 循环学习率10)Learning Rate Schedule - 学习率调度11)Warm-up - 预热12)Learning Rate Policy - 学习率策略13)Learning Rate Annealing - 学习率退火14)Cosine Annealing - 余弦退火15)Gradient Clipping - 梯度裁剪16)Adapting Learning Rate - 适应学习率17)Learning Rate Multiplier - 学习率倍增器18)Learning Rate Reduction - 学习率降低19)Learning Rate Update - 学习率更新20)Scheduled Learning Rate - 定期学习率•Batch Size - 批量大小1)Batch Size - 批量大小2)Mini-Batch - 小批量3)Batch Gradient Descent - 批量梯度下降4)Stochastic Gradient Descent (SGD) - 随机梯度下降5)Mini-Batch Gradient Descent - 小批量梯度下降6)Online Learning - 在线学习7)Full-Batch - 全批量8)Data Batch - 数据批次9)Training Batch - 训练批次10)Batch Normalization - 批量归一化11)Batch-wise Optimization - 批量优化12)Batch Processing - 批量处理13)Batch Sampling - 批量采样14)Adaptive Batch Size - 自适应批量大小15)Batch Splitting - 批量分割16)Dynamic Batch Size - 动态批量大小17)Fixed Batch Size - 固定批量大小18)Batch-wise Inference - 批量推理19)Batch-wise Training - 批量训练20)Batch Shuffling - 批量洗牌•Epoch - 训练周期1)Training Epoch - 训练周期2)Epoch Size - 周期大小3)Early Stopping - 提前停止4)Validation Set - 验证集5)Training Set - 训练集6)Test Set - 测试集7)Overfitting - 过拟合8)Underfitting - 欠拟合9)Model Evaluation - 模型评估10)Model Selection - 模型选择11)Hyperparameter Tuning - 超参数调优12)Cross-Validation - 交叉验证13)K-fold Cross-Validation - K折交叉验证14)Stratified Cross-Validation - 分层交叉验证15)Leave-One-Out Cross-Validation (LOOCV) - 留一法交叉验证16)Grid Search - 网格搜索17)Random Search - 随机搜索18)Model Complexity - 模型复杂度19)Learning Curve - 学习曲线20)Convergence - 收敛3.Machine Learning Techniques and Algorithms (机器学习技术与算法)•Decision Tree - 决策树1)Decision Tree - 决策树2)Node - 节点3)Root Node - 根节点4)Leaf Node - 叶节点5)Internal Node - 内部节点6)Splitting Criterion - 分裂准则7)Gini Impurity - 基尼不纯度8)Entropy - 熵9)Information Gain - 信息增益10)Gain Ratio - 增益率11)Pruning - 剪枝12)Recursive Partitioning - 递归分割13)CART (Classification and Regression Trees) - 分类回归树14)ID3 (Iterative Dichotomiser 3) - 迭代二叉树315)C4.5 (successor of ID3) - C4.5（ID3的后继者）16)C5.0 (successor of C4.5) - C5.0（C4.5的后继者）17)Split Point - 分裂点18)Decision Boundary - 决策边界19)Pruned Tree - 剪枝后的树20)Decision Tree Ensemble - 决策树集成•Random Forest - 随机森林1)Random Forest - 随机森林2)Ensemble Learning - 集成学习3)Bootstrap Sampling - 自助采样4)Bagging (Bootstrap Aggregating) - 装袋法5)Out-of-Bag (OOB) Error - 袋外误差6)Feature Subset - 特征子集7)Decision Tree - 决策树8)Base Estimator - 基础估计器9)Tree Depth - 树深度10)Randomization - 随机化11)Majority Voting - 多数投票12)Feature Importance - 特征重要性13)OOB Score - 袋外得分14)Forest Size - 森林大小15)Max Features - 最大特征数16)Min Samples Split - 最小分裂样本数17)Min Samples Leaf - 最小叶节点样本数18)Gini Impurity - 基尼不纯度19)Entropy - 熵20)Variable Importance - 变量重要性•Support Vector Machine (SVM) - 支持向量机1)Support Vector Machine (SVM) - 支持向量机2)Hyperplane - 超平面3)Kernel Trick - 核技巧4)Kernel Function - 核函数5)Margin - 间隔6)Support Vectors - 支持向量7)Decision Boundary - 决策边界8)Maximum Margin Classifier - 最大间隔分类器9)Soft Margin Classifier - 软间隔分类器10) C Parameter - C参数11)Radial Basis Function (RBF) Kernel - 径向基函数核12)Polynomial Kernel - 多项式核13)Linear Kernel - 线性核14)Quadratic Kernel - 二次核15)Gaussian Kernel - 高斯核16)Regularization - 正则化17)Dual Problem - 对偶问题18)Primal Problem - 原始问题19)Kernelized SVM - 核化支持向量机20)Multiclass SVM - 多类支持向量机•K-Nearest Neighbors (KNN) - K-最近邻1)K-Nearest Neighbors (KNN) - K-最近邻2)Nearest Neighbor - 最近邻3)Distance Metric - 距离度量4)Euclidean Distance - 欧氏距离5)Manhattan Distance - 曼哈顿距离6)Minkowski Distance - 闵可夫斯基距离7)Cosine Similarity - 余弦相似度8)K Value - K值9)Majority Voting - 多数投票10)Weighted KNN - 加权KNN11)Radius Neighbors - 半径邻居12)Ball Tree - 球树13)KD Tree - KD树14)Locality-Sensitive Hashing (LSH) - 局部敏感哈希15)Curse of Dimensionality - 维度灾难16)Class Label - 类标签17)Training Set - 训练集18)Test Set - 测试集19)Validation Set - 验证集20)Cross-Validation - 交叉验证•Naive Bayes - 朴素贝叶斯1)Naive Bayes - 朴素贝叶斯2)Bayes' Theorem - 贝叶斯定理3)Prior Probability - 先验概率4)Posterior Probability - 后验概率5)Likelihood - 似然6)Class Conditional Probability - 类条件概率7)Feature Independence Assumption - 特征独立假设8)Multinomial Naive Bayes - 多项式朴素贝叶斯9)Gaussian Naive Bayes - 高斯朴素贝叶斯10)Bernoulli Naive Bayes - 伯努利朴素贝叶斯11)Laplace Smoothing - 拉普拉斯平滑12)Add-One Smoothing - 加一平滑13)Maximum A Posteriori (MAP) - 最大后验概率14)Maximum Likelihood Estimation (MLE) - 最大似然估计15)Classification - 分类16)Feature Vectors - 特征向量17)Training Set - 训练集18)Test Set - 测试集19)Class Label - 类标签20)Confusion Matrix - 混淆矩阵•Clustering - 聚类1)Clustering - 聚类2)Centroid - 质心3)Cluster Analysis - 聚类分析4)Partitioning Clustering - 划分式聚类5)Hierarchical Clustering - 层次聚类6)Density-Based Clustering - 基于密度的聚类7)K-Means Clustering - K均值聚类8)K-Medoids Clustering - K中心点聚类9)DBSCAN (Density-Based Spatial Clustering of Applications with Noise) - 基于密度的空间聚类算法10)Agglomerative Clustering - 聚合式聚类11)Dendrogram - 系统树图12)Silhouette Score - 轮廓系数13)Elbow Method - 肘部法则14)Clustering Validation - 聚类验证15)Intra-cluster Distance - 类内距离16)Inter-cluster Distance - 类间距离17)Cluster Cohesion - 类内连贯性18)Cluster Separation - 类间分离度19)Cluster Assignment - 聚类分配20)Cluster Label - 聚类标签•K-Means - K-均值1)K-Means - K-均值2)Centroid - 质心3)Cluster - 聚类4)Cluster Center - 聚类中心5)Cluster Assignment - 聚类分配6)Cluster Analysis - 聚类分析7)K Value - K值8)Elbow Method - 肘部法则9)Inertia - 惯性10)Silhouette Score - 轮廓系数11)Convergence - 收敛12)Initialization - 初始化13)Euclidean Distance - 欧氏距离14)Manhattan Distance - 曼哈顿距离15)Distance Metric - 距离度量16)Cluster Radius - 聚类半径17)Within-Cluster Variation - 类内变异18)Cluster Quality - 聚类质量19)Clustering Algorithm - 聚类算法20)Clustering Validation - 聚类验证•Dimensionality Reduction - 降维1)Dimensionality Reduction - 降维2)Feature Extraction - 特征提取3)Feature Selection - 特征选择4)Principal Component Analysis (PCA) - 主成分分析5)Singular Value Decomposition (SVD) - 奇异值分解6)Linear Discriminant Analysis (LDA) - 线性判别分析7)t-Distributed Stochastic Neighbor Embedding (t-SNE) - t-分布随机邻域嵌入8)Autoencoder - 自编码器9)Manifold Learning - 流形学习10)Locally Linear Embedding (LLE) - 局部线性嵌入11)Isomap - 等度量映射12)Uniform Manifold Approximation and Projection (UMAP) - 均匀流形逼近与投影13)Kernel PCA - 核主成分分析14)Non-negative Matrix Factorization (NMF) - 非负矩阵分解15)Independent Component Analysis (ICA) - 独立成分分析16)Variational Autoencoder (VAE) - 变分自编码器17)Sparse Coding - 稀疏编码18)Random Projection - 随机投影19)Neighborhood Preserving Embedding (NPE) - 保持邻域结构的嵌入20)Curvilinear Component Analysis (CCA) - 曲线成分分析•Principal Component Analysis (PCA) - 主成分分析1)Principal Component Analysis (PCA) - 主成分分析2)Eigenvector - 特征向量3)Eigenvalue - 特征值4)Covariance Matrix - 协方差矩阵。

通信专业英文缩略语缩略语英语解释中文解释AAM/CM Administration Module/Communication Module 管理和通信模块BBA BCCH Allocation BCCH分配BAM Back Administration Module 后管理模块BCC BTS Color Code 基站色码BCCH Broadcast Control CHannel 播送控制信道BCH Broadcast channel (transport channel) 播送信道〔传输信道〕BIE Base Station Interface Equipment 基站接口设备BITS Building Integrated Timing Supply 大楼综合定时供应系统BM Basic Module 根本模块BS Base Station 基站BS1 Abis Interface Abis接口BSC Base Station Controller 基站控制器BSIU Base Station Interface Unit 基站接口单元BSS Base Station Subsystem 基站子系统BTS Base Transceiver Station 基站收发信台CCA Cell Allocation 小区分配CBCH Cell Broadcast Channel 小区播送信道CC Country Code 国家码CCCH Common Control Channel 公共控制信道CCS Common Channel Signaling 共路信令方式CDU Combiner and Divider Unit 合分路单元CGI Cell Global Identification 全球小区识别码CIC Circuit Identification Code 电路识别码CRC Cyclic Redundancy Check 循环冗余校验CS-1 Code Scheme-1 编码模式－1〔〕CS-2 Code Scheme-2 编码模式－2〔〕CS-3 Code Scheme-3 编码模式－3〔〕CS-4 Code Scheme-4 编码模式－4〔〕CTN Central Switching Network Board 中央交换网板DDBF Database File 数据库文件DPC Destination (Signaling) Point Code 目的信令点编码DRX Discontinuous Reception 非连续接收DTX Discontinuous Transmission 非连续性发射EE3M E3 Sub-Multiplexer E3子复用设备EAC External Alarm Collection 外部告警采集ECSC Early Classmark Sending Control 早期类标发送控制EDU Enhanced Duplexer Unit 增强型双工单元EFR Enhanced full rate speech code 增强型全速率语音编解码EST Establishment 建立FFACCH Fast Associated Control Channel 快速随路控制信道FBC Photoelectric Conversion Board 光电转换板FBI Optical Fiber Interface Board 光接口板FPU Frame Processing Unit 帧处理单元FTC Full Rate Transcoder 全速率码变换器、全速率变码器FTP File Transfer Protocol 文件传输协议FUL Radio Signaling Link 无线信令链路GGALM Alarm board 告警板GCKS Clock source 时钟板GCTN Central switching Network board 中心交换网板GEMA Emergency Action board 双机倒换板GFBI Fiber Interface board 光纤接口板GLAP LAPD Protocol Process board LAPD协议处理板GMC2 Inter-Module Communication board，模块通信板GMCC Module Communication and Control board 模块通信控制板GMEM Memory board 数据库接口板GMPU Main Process Unit 主处理单元GNET Switching Network board 交换网板GNOD Node Communication Board 节点通信板GOPT Local Optical Interface Board 光纤通信板GPRS General Packet Radio Service 通用分组无线业务GPS Global Position System 全球定位系统GSM Global System for Mobile Communications 全球移动通信系统GSNT GSM Signaling Switching Network Board 信令交换网板HHDLC High-level Data Link Control 高级数据链路控制HO Handover 切换HPA High magnification Power Amplifier board 高增益功放板HSN Hopping Sequence Number 跳频序列号HW Highway 母线IID IDentification/IDentity 识别IND Indication 指示IOMU iSite Operation and Maintenance Unit 操作维护单元板IP Internet Protocol 互联网协议、网际协议LLAPD Link Access Protocol on the D-channel D信道上的链路访问协议LPN7 Common Channel Signaling Processing Board 公共信道信令处理板MMA Mobile Allocation 移动台〔频率〕分配MAIO Mobile Allocation Index Offset 移动分配索引偏移MCC Mobile Country Code 移动国家码MCK Main Clock board 主时钟板MFU Microcell Frame Unit 微蜂窝帧处理单元MMU Multiplication and Management Unit 复用管理单元MNC Mobile Network Code 移动网编号、移动网编码MS Mobile Station 移动台()MSC Mobile Switching Center 移动交换中心MSM MSC Subrate channel Multiplexer MSC侧子复用板MTP Message Transfer Part 消息传递局部NNCC Network Color Code 网络色码NSS Network SubSystem 网络子系统OOM Opration and Maintenance 操作维护OMC Operation and MaintenanceCenter 操作维护中心OML Operation and Maintenance Link 操作与维护链路OMU Operation and Maintenance Unit 操作维护单元OPC Originating Point Code 源信令点编码PPb Pb Interface Pb接口PBCCH Packet Broadcast Control Channel 分组播送控制信道PBGT Power Budget 功率预算PBU Power Boost Unit 功率增强单元PCCCH Packet Common Control Channel 分组公共控制信道PCIC Packet Circuit Identity Code 分组电路标识码PCM Pulse-Code Modulation 脉冲编码调制PCU Packet Control Unit 分组控制单元PDH Plesiochronous Digital Hierarchy 准同步数字系列PDTCH Packet Data Traffic Channel 分组业务数据信道PLMN Public Land Mobile Network 公用陆地移动〔通信〕网PMU Power and Environment Monitoring Unit 电源环境监测板PSU Power Supply Unit 供电单元PWC Secondary Power Supply Board 二次电源板RRACH Random Access CHannel 随机接入信道RSL Radio Signaling Link 无线信令链路SSACCH Slow Associated Control Channel 慢速随路控制信道SAPI Service Access Point Identifier 业务接入点标识SCCP Signaling Connection Control Part 信令连接控制局部SCU Simple Combiner Unit 简单合路单元SDCCH Stand-alone Dedicated Control Channel 独立专用控制信道SITE Site 站点SM Sub-Multiplexer Interface 子复用板SMBCB Short Message Service Cell Broadcast 短消息业务小区播送SMI Sub-Multiplexer Interface 子复用板SP Signaling Point 信令点SS7 Signaling System Number 7 七号信令STP Signaling Transfer Point 信令转接点TTA Timing Advance 时间提前量TC Transcoder 码变换器TCH Traffic CHannel 业务信道TCSM Transcoder and Sub-Multiplexer 码变换与子复用单元〔器〕TEI Terminal Equipment Identifier 终端设备标识TES Transmission Extension power Supply unit 传输扩展供电单元TEU Transmission Extension Unit 传输扩展单元TFO Tandem Free Operation 免汇接运营TMU Timing/Transmission and Management Unit 定时/传输管理单元TRAU Transcoder & Rate Adaptation Unit 码变换器/速率适配单元TRX Transceiver 收发信机TS Timeslot 时隙TSC Training Sequence Code 训练系列号(编码〕VVSWR Voltage Standing Wave Ratio 电压驻波比WWS Workstation 操作台A, Asub A-interface A接口AC Alternating Current 交流AC Access Class (C0 to C15) 接入级别〔C0到C15〕ACCH Associated Control Channel 随路控制信道ACELP Algebraic code excitation linear prediction 代数码鼓励线性预测ACOM Antenna Combiner 天线合路器AGCH Access Grant Channel 接入允许信道AM/CM Administration Module/ Communication Module 管理和通信模块ANSI American National Standard Institute 美国国家标准组织APC Automatic Power Control 自动功率控制API Application Program Interface 应用程序接口APL Advanced Phase Locking 高级时钟锁相ARFCN Absolute Radio Frequency Channel Number 绝对射频信道号ASIC Application Specific Integrated Circuit 专用集成电路AuC Authentication Center 鉴权中心BBA BCCH Allocation BCCH分配BAM Back Administration Module 后管理模块BCC BTS Color Code 基站色码BCCH Broadcast Control CHannel 播送控制信道BCF Base Control Function 根本控制功能BCH Broadcast channel (transport channel) 播送信道BER Bit Error Rate 误码率BHCA Busy Hour Call Attempt 忙时尝试呼叫BIE Base station Interface Equipment (board) 基站接口设备〔板〕BIOS Basic Input Output System 根本输入输出系统BITS Building Integrated Timing Supply 大楼综合定时供应系统BM Basic Module 根本模块BP Burst Pulse 突发脉冲BQ Bad Quality 质量差BS Base Station 基站BS1 Abis Interface Abis接口BSC Base Station Controller 基站控制器BSIC Base Station Identity Code 基站识别码BSMU Base Station Interface Unit 基站接口单元BSS Base Station Subsystem 基站子系统BSSAP Base Station Subsystem Application Part 基站子系统应用局部BSSGP Base Station Subsystem GPRS Protocol 基站系统GPRS协议BSSMAP Base Station Subsystem Management Application Part 基站子系统管理应用局部BSSOMAP Base Station Subsystem Operation and Maintenance Application Part 基站子系统操作与维护应用局部BTS Base Transceiver Station 基站收发信台BTSM Base Transceiver Station Management BTS管理BVC BSSGP Virtual Connection BSSGP虚拟连接CCA Cell Allocation 小区分配CAMEL Customized Applications for Mobile network Enhanced Logic 移动网络增强逻辑的客户化应用CBA Cell Bar Access 小区禁止接入CBC Cell Broadcast Center 小区播送中心CBCH Cell Broadcast CHannel 小区播送信道CBCCH Cell Broadcast Control Channel 小区播送控制信道CBQ Cell Bar Qualify 小区禁止限制CBSM Cell Broadcast Short Message 小区播送短消息CC Country Code 国家码CC Calling Control 呼叫控制CC Connection Confirm 呼叫控制CCB Call Control Block 呼叫控制块CCBS Completion of Calls to Busy Subscribers 遇忙回呼CCCH Common Control Channel 公共控制信道CCH Control Channel 控制信道CCS Common Channel Signaling 共路信令方式CD Call Deflection 呼叫偏移CDB Cell Broadcast Database 小区播送数据库CDU Combining and Distribution Unit 合分路单元CELP Code Excited Linear Prediction 码鼓励线性预测CGI Cell Global Identity 小区全球识别码CI Cell Identity 小区识别CIC Circuit Identify Code 电路识别码CIC Carrier Interface Controller board 载频接口控制器CIR Carrier to Interference Ratio 载干比CKSN Ciphering Key Sequence Number 密钥序列号CKV Clock Drive board 时钟驱动板CM Connection Management 接续管理CPU Central Processing Unit 中央处理单元CR Connection Request 连接请求CRC Cyclic Redundancy Check 循环冗余校验CRO Cell Reselect Offset 小区重选偏移CS Coding Scheme 〔信道〕编码方式CS-1 Code Scheme-1 编码模式－1〔9.05kbit/s〕CS-2 Code Scheme-2 编码模式－2〔13.4kbit/s〕CS-3 Code Scheme-3 编码模式－3〔15.6kbit/s〕CS-4 Code Scheme-4 编码模式－4〔21.4kbit/s〕CTN Central Switching Network Board 中央交换网板DDB DataBase 数据库DBF Database File 数据库文件DBMS Database Management System 数据库管理系统DC Direct Current 直流DCCH Dedicated Control Channel 专用控制信道DCL Diagnostic Control Link 诊断控制链路DDN Digital Data Network 数字数据网DL Downlink 下行链路DLC Data Link Connection 数据链路连接DLCEP Data Link Connection End Point 数据链路连接端点DLCEPI Data Link Connection End Point Identifier 数据链路连接端点标识DLCI Digital Link Connection Identity 数据链路连接标识DNS Domain Name Server 域名效劳器DPC Destination (Signaling) Point Code 目的信令点编码DRDBMS Distributed Relational DBMS 分布式关系数据库管理系统DRX Discontinuous Reception (mechanism) 不连续接收DSC Downlink Signaling fault Count 下行信令故障计数DSP Digital Signal Processor 数字信号处理器DTAP Direct Transfer Application Part 直接传输应用局部DTMF Dual Tone Multi-frequency 双音多频〔收号器〕DTX Discontinuous transmission (mechanism) 不连续发送〔机制〕EE-Abis Enhanced Abis 增强型AbisE3M E3 Sub-Multiplexer 增强型E1子复用设备EA Early Allocation 预分配EAC External Alarm Collection 外部告警采集EC Emergency Call 紧急呼叫ECSC Early Classmark Sending Control 早期类标发送控制ECT Explicit Call Transfer 显示呼叫转移EDU Enhanced Duplexer Unit 增强型双工单元EFR Enhanced full rate speech code 增强型全速率语音编解码EIR Equipment Identity Register 设备识别存放器EM Extended Measurement 扩展测量EMC Electromagnetic Compatibility 电磁兼容性EST Establishment 建立ETS European Telecommunication Standard 欧洲电信标准ETSI European Telecommunication Standard Institute 欧洲电信标准组织FFACCH Fast Associated Control CHannel 快速随路控制信道FBC Photoelectric Conversion Board 光电转换板FBI Optical Fiber Interface Board 光接口板FCCH Frequency Correction CHannel 频率校正信道FCS Frame Check Sequence 帧校验序列FDMA Frequency Division Multiple Access 频分多址FH Frequency Hopping 跳频FIR Finity Impulsion Response 有限冲击响应FN Frame Number 帧号FPU Frame Processing Unit 帧处理单元FR Frame Relay 帧中继FTAM File Transfer Access and manipulation 文件传输、接入及使用FTC Full Rate Transcoder 码变换板FTP File Transfer Protocol 全速率码变换器FUC Frame Unit Controller 帧单元控制器FUL Radio Signaling Link 无线信令链路GG-Abis GPRS Abis GPRS AbisGALM Alarm board 告警板GCKS Clock source 时钟板GCTN Central switching Network board 中心交换网板GEMA Emergency Message Automatic Transmission System 双机倒换板GFBI Fiber Interface board 光纤接口板GGSN Gateway GPRS Support Node 网关GPRS支持节点GLAP LAPD Protocol Process board LAPD协议处理板GMC2 Inter-Module Communication board 模块通信板GMCC Module Communication and Control board 模块通信控制板GMEM Memory board 数据库接口板GMM GPRS Mobility Management GPRS移动性管理GMPU Main Processing Unit 主处理单元GMSC Gateway Mobile Switching Center 关口局GMSK Gaussian Minimum Shift-frequency Keying 高斯滤波最小移频键控GNET Intra-module switching network board 交换网板GNOD Node Communication Board 节点通信板GOPT Local Optical Interface Board 光纤通信板GPRS General Packet Radio Service 通用分组无线业务GPS Global Position System 全球定位系统GPWS GSM Secondary Power board 二次电源板GSM，GSM900，GSM1800 Global System for Mobile communications 全球移动通信系统，900MHz的GSM系统，1800MHz的GSM系统GSN GPRS Support Node GPRS支持节点GSNT GSM Signaling Switching Network Board 信令交换网板GT Global Title 全局码GTP GPRS Tunnelling Protocol GPRS隧道协议HHC/HY COM Hybrid Combiner 混合桥型合路器HCS Hierarchical Cell Structure 小区分层结构HDLC High level Data Link Control 高级数据链路控制HDSL High speed Digital Subscriber Line 高速数字用户线HLR Home Location Register 归属位置存放器HO Handover 切换HPA High magnification Power Amplifier board 高增益功放板HSC Hot Swap Controller 热倒换控制器HSN Hopping Sequence Number 跳频序列号HW Highway 高速通路IID IDentification/IDentity 识别IEC International Electrotechnical Commission 国际电工委员会IMEI International Mobile station Equipment Identity 国际移动终端设备标识IMSI International Mobile Station Identity 国际移动用户识别码IND Indication 指示IOMU iSite Operation and Maintenance Unit 操作维护单元板IP Internet Protocol 互联网协议ISDN Integrated Services Digital Network 综合业务数字网ISO International Standard Organization 国际标准化组织ISR Interrupt Service 中断效劳程序ISUP Integrated Services Digital Network User Part/ISDN User Part 〔七号信令之〕ISDN用户局部ITU International Telecommunication Union 国际电信联盟ITU-T International Telecommunication Union - Telecommunication Standardization Sector 国际电信联盟-电信标准部IWF Inter-working Function 互连功能J- -KLL2ML Layer 2 Management Link 层2管理链路L3MM Layer-3 Mobility Management 层三移动管理LA Location Area 位置区LAC Location Area Code 位置区码〔LAC〕LAI Location Area Identity 位置区标识LAP Link Access Protocol 协议处理板LAPD Link Access Protocol on the D-channel D信道上的链路访问协议LAPDMAIL LAPD Mail Box LAPD邮箱LAPDm Link Access Protocol on the Dm channel Dm信道上的链路访问协议LLC Logical Link Control 逻辑链路控制LMT Local Maintenance Terminal 本地维护终端LNA Low Noise Amplifier 低噪声放大器LPN7 Common Channel Signaling Processing Board 公共信道信令处理板MMA Mobile Allocation 移动台〔频率〕分配MAC Media Access Control 媒质接入控制MAIO Mobile Allocation Index Offset 移动分配索引偏移MAP Mobile Application Part 移动应用局部MBR Multiband Report 多频报告MCC Mobile Country Code 移动国家码MCK Main ClocK board 主时钟板MCP Multiple Communication-Protocol Processor 多重通信协议处理器MDC Message Discrimination 消息鉴别MDSL Medium Bit-rate Digital Subscriber Loop 中速数字用户环线MDT Message Distribution 消息分配ME Mobile Equipment 移动设备MFU Microcell Frame Unit 微蜂窝帧处理单元MM Mobility Management 移动性管理MMU Multiplication and Management Unit 复用管理单元MNC Mobile Network Code 移动网号MNS Mobile Network Signaling 移动网信令MR Measurement Result 测量结果MR Measurement Report 测量报告MRP Multiple Reuse Pattern 多重复用方式MRT Message Routing 消息路由MS Mobile Station 测量报告MSC Mobile services Switching Centre, Mobile Switching Centre 移动交换中心MSISDN Mobile Station International ISDN Number 移动台国际ISDN号码MSM MSC Subrate channel Multiplexer MSC侧子复用板MT Mobile Terminal 移动终端MTBF Mean Time Between Failure 平均无故障时间MTP Message Transfer Part 消息传输局部NNC Network Control 网络控制NCC Network Color Code 网络色码NCH Notification Channel 通知信道NE Network Equipment 网络设备NM Network Management 网络管理NS Network Service 网络效劳NSE Network Service Entity 网络效劳实体NSS Network SubSystem 网络子系统OO&M, OM Operations & Maintenance 操作与维护OACSO Off Air Call Set up 不占用空中通道的呼叫启动OAM Operation Administration and Maintenance 运行管理和维护OMAP Operation and Maintenance Application Part 操作维护应用局部OMC Operations & Maintenance Centre 操作维护中心OML Operation and Maintenance Link 操作与维护链路OMU Operations & Maintenance Unit (board) 操作维护单元〔板〕OOP Object Oriented Programming 面向对象的程序设计OPC Originating Point Code 源信令点编码OPT Optic Interface board 光纤通信板OS Operation System 操作系统OSI Open System Interconnection 开放系统互连模型PPA Power Amplifier 功率放大器PAGCH Packet Access Grant Channel 分组接入允许信道PBCCH Packet Broadcast Control Channel 分组播送控制信道PBGT Power Budget 功率预算PBU Power Boost Unit 功率增强单元Pb Pb Interface Pb接口PbSL PCU-BSC Signaling Link PCU-BSC间信令链路PCCCH Packet Common Control Channel 分组公共控制信道PCH Paging CHannel 寻呼信道PCIC Packet Circuit Identity Code 分组电路标识码PCM Pulse-Code Modulation 脉冲编码调制PCU Packet Control Unit 分组控制单元PD Protocol Discrimination 协议识别码PDCH Packet Data Channel 分组数据信道PDH Plesiochronous Digital Hierarchy 准同步数字系列PDN Packet Data Network 分组数据网PDP Packet Data Protocol 分组数据协议PDTCH Packet Data Traffic Channel 分组业务数据信道PI Peripheral Interface 外设接口部件PIN Personal Identity Number 个人识别码PLL Phase Locked Loop 锁相环路PLMN Public Land Mobile Network 公用陆地移动网络PMU Power and Environment Monitoring Unit 电源环境监测板PNCH Packet Notification Channel 分组通知信道POMU Packet Operation & Maintenance Unit 分组操作维护单元PON Passive Optical Network 无源光网络PPCH Packet Paging Channel 分组寻呼信道PRACH Packet Random Access Channel 分组随机接入信道PSDN Public Switched Data Network 公用数据交换网PSI Packet System Information 分组系统消息PSK Phase Shift Keying 相移键控PSTN Public Switched Telephone Network 公用交换网PSU Power Supply Unit 供电单元PT Penalty Time 惩罚时间PTCCH Packet Timing advance Control Channel 分组定时提前控制信道PTM Point To Multipoint 点到多点PTM-M Point To Multipoint Multicast 点对多点播送PTM-SC Point to Multipoint Service Center 点到多点数据效劳中心PTP Point To Point 点对点PWC Secondary Power Supply Board 电源控制板QQoS Quality of Service 业务质量RRACH Random Access Channel 随机接入信道RE Reestablishment 呼叫重建RF Radio Frequency 射频RLC Radio Link Control 无线链路控制RLM Radio Link Management 无线链路管理RPE-LTP Regular Pulse Excitation-Long Term Prediction 规那么脉冲鼓励-长期预测RPPU Radio Packet Process Unit 无线分组处理单元RR Radio Resource 无线资源RSA Rivest-Shamir-Adleman 通用关键子密码方法RSL Radio Signaling Link 无线信令链路RTE Radio Test Equipment 天线测试设备RX Receiver/Reception 收信机/接收RXLEV Received signal level 接收信号等级RXQUAL Received Signal Quality 接收信号质量SSABM Set Asynchronous Balanced Mode 置异步平衡模式SACCH Slow Associated Control Channel 慢速随路控制信道SACCH/C4 Slow Associated Control Channel/SDCCH/4 慢速随路控制信道/SDCCH/4 SACCH/C8 Slow Associated Control CHannel/SDCCH/8 慢速随路控制信道/SDCCH/8 SACCH/T Slow Associated Control CHannel/Traffic channel 慢速随路控制信道/业务信道SACCH/TF Slow Associated Control Channel/Traffic channel Full rate 慢速随路控制信道/全速率业务信道SAP Service Access Point 效劳接入点SAPI Service Access Point Identifier 业务接入点标识SCCP Signaling Connection Control Part 信令连接控制局部SCH Synchronization CHannel 同步信道SCMG SCCP Management SCCP管理SCU Simple combining Unit 简单合路单元SDCCH Stand-alone Dedicated Control CHannel 独立专用控制信道SDH Synchronous Digital Hierarchy 同步数字系列SDU Service Data Unit 业务数据单元SGSN Serving GPRS Support Node 效劳GPRS支持节点SID Silence Descriptor 静噪指示SIG Signaling 信令SIM Subscriber Identity Module 用户识别卡SITE Site 站点SLM Signaling Link Management 信令链路管理SLS Signaling Link Selection 信令链路选择码SM Sub-Multiplexer Interface 子复用板SM-SC Short Message - Service Center 短消息中心SMBCB Short Message Service Cell Broadcast 短消息业务小区播送SMC Short Message Center 短消息中心SMI Sub-Multiplexer Interface 子复用板SMS Short Message Service 短消息业务SMS-GMSC Short Message Service - Gateway MSC 短消息关口MSC SMS-IWMSC Short Message Service Interworking MSC 短消息互联MSC SMSCB Short Message Service Cell Broadcast 短消息小区播送SMUX Sub-Multiplexer 子复用器SNDCP SubNetwork Dependent convergence Protocol 子网相关的收敛协议SOR Support Optimization Routing 支持优选路由SP Signaling Point 信令点SRM Signaling Route Management 信令路由管理SS Supplementary Service 补充业务SS7 Signalling System No.7 七号信令SSN SubSystem Number 子系统号STM Signaling Traffic Management 信令业务管理STP Signaling Transfer Point 信令转接点TTA Timing Advance 定时提前量TAI Timing Advance Index 时间提前量索引TBF Temporary Block Flow 临时数据块流TC Transcoder 码变换器TCH Traffic Channel 业务信道TCH/F A full rate TCH 全速率业务信道TCH/F A full rate data TCH (2.4kbit/s) 全速率数据业务信道〔2.4kbit/s〕TCH/F A full rate date TCH (4.8kbit/s) 全速率数据业务信道〔4.8kbit/s〕TCH/F A full rate data TCH (9.6kbit/s) 全速率数据业务信道〔9.6kbit/s〕TCH/FS A full rate Speech TCH 全速率话音业务信道TCI Terminal Interface board 终端接口板TCP Transmission Control Protocol 传输控制协议TCSM TransCoder & Sub-Multiplexer 码变换与子复用单元TDMA Time Division Multiple Access 时分多址TE Terminal Equipment 终端设备TEI Terminal Equipment Identifier 终端设备识别码TES Transmission Extension power Supply unit 传输扩展供电单元TEU Transmission Extension Unit 传输扩展单元TFI Transport Format Indicator 传输格式指示TFO Tandem Free Operation 免汇接运营TLLI Temporary Link Level Identity 临时链路等级标识TMSC Tandem Mobile Switching Centre 汇接移动交换中心TMSI Temporary Mobile Subscriber Identifier 临时移动用户标识符TMU Timing/Transmission and Management Unit 定时/传输管理单元TN Timeslot Number 时隙号TNI Terminal Network Interface 终端网络接口部件TO Temporary Offset 临时偏移TRAU Transcoder & Rate Adaptation Unit 码变换器/速率适配单元TRX Transceiver (board) 收发信机TS Timeslot 时隙TSC Training Sequence Code 训练系列号(编码〕TUP Telephone User Part(SS7) 用户局部UUA Unnumbered Acknowledge 无编号证实UDP User Datagram Protocol 用户数据报协议UDT Unit Data 单位数据UI Unnumbered Information (frame) 无编号信息帧Um 空中接口USF Uplink State Flag 上行链路状态标识USSD Unstructured Supplementary Service Data 非结构化补充业务数据VV AD V oice Activity Detection 话音激活检测VBS V oice Broadcast Service 话音播送呼叫业务VEA Very Early Allocation 很早分配VGCS V oice Group Call Service 话音组呼业务VLR Visitor Location Register 拜访用户位置存放器VM Voice Mailbox 语音邮箱VSAT Very Small Aperture Terminal 甚小天线卫星地球站WWDT Watchdog Timer 看门狗WS Workstation 操作台XxDSL x Digital Subscriber Line x数字用户线YZ爱尔兰表：某基站站型为S4/4/4。

模拟ai英文面试题目及答案模拟AI英文面试题目及答案1. 题目: What is the difference between a neural network anda deep learning model?答案: A neural network is a set of algorithms modeled loosely after the human brain that are designed to recognize patterns. A deep learning model is a neural network with multiple layers, allowing it to learn more complex patterns and features from data.2. 题目: Explain the concept of 'overfitting' in machine learning.答案: Overfitting occurs when a machine learning model learns the training data too well, including its noise and outliers, resulting in poor generalization to new, unseen data.3. 题目: What is the role of a 'bias' in an AI model?答案: Bias in an AI model refers to the systematic errors introduced by the model during the learning process. It can be due to the choice of model, the training data, or the algorithm's assumptions, and it can lead to unfair or inaccurate predictions.4. 题目: Describe the importance of data preprocessing in AI.答案: Data preprocessing is crucial in AI as it involves cleaning, transforming, and reducing the data to a suitableformat for the model to learn effectively. Proper preprocessing can significantly improve the performance of AI models by ensuring that the input data is relevant, accurate, and free from noise.5. 题目: How does reinforcement learning differ from supervised learning?答案: Reinforcement learning is a type of machine learning where an agent learns to make decisions by performing actions in an environment to maximize a reward signal. It differs from supervised learning, where the model learns from labeled data to predict outcomes based on input features.6. 题目: What is the purpose of a 'convolutional neural network' (CNN)?答案: A convolutional neural network (CNN) is a type of deep learning model that is particularly effective for processing data with a grid-like topology, such as images. CNNs use convolutional layers to automatically and adaptively learn spatial hierarchies of features from input images.7. 题目: Explain the concept of 'feature extraction' in AI.答案: Feature extraction in AI is the process of identifying and extracting relevant pieces of information from the raw data. It is a crucial step in many machine learning algorithms, as it helps to reduce the dimensionality of the data and to focus on the most informative aspects that can be used to make predictions or classifications.8. 题目: What is the significance of 'gradient descent' in training AI models?答案: Gradient descent is an optimization algorithm used to minimize a function by iteratively moving in the direction of steepest descent as defined by the negative of the gradient. In the context of AI, it is used to minimize the loss function of a model, thus refining the model's parameters to improve its accuracy.9. 题目: How does 'transfer learning' work in AI?答案: Transfer learning is a technique where a pre-trained model is used as the starting point for learning a new task. It leverages the knowledge gained from one problem to improve performance on a different but related problem, reducing the need for large amounts of labeled data and computational resources.10. 题目: What is the role of 'regularization' in preventing overfitting?答案: Regularization is a technique used to prevent overfitting by adding a penalty term to the loss function, which discourages overly complex models. It helps to control the model's capacity, forcing it to generalize better to new data by not fitting too closely to the training data.。

上海科技馆导游讲解词eps down on both sides, then we can go into science and technology museum size 6 door.6号门入口处就有购票点，但是如果你错过了也不用担心，由电梯上到一楼也可购票。

6# door entrance has tickets site, but if you missed also need not worry, by the elevator up to the first floor can also be tickets.上楼后，在你的后方就是科技馆的检票口了。

先别急，别忘了拿小架子上的免费地图和展馆介绍，要不然你准会迷路。

要知道对着地图和介绍，合理安排好自己的时间和行程还是十分必要的，磨刀不误砍柴功哦。

Upstairs, your rear is science and technology museum in the ticket. B: don't worry, don't forget to take small shelf free maps and e某hibition hall, otherwise you'll introduce got lost. Want to know the map and introduce to arrange their own time and travel or very necessary, sharpening will not delay your job of work well.下面就跟着我一起领略一下各有特色的展馆吧!Below follow me appreciating the have distinguishing feature each e某hibition!进门后在你的左边有一个巨大的绿色王国，这就是“生物万象”展区。

Using Agents for Distributed Software Project Management Rory O’Connor John Jenkins School of Computer Applications School of Computing Science Dublin City University Middlesex University Ireland UK roconnor@compapp.dcu.ieAbstractThe paper explores the role of artificial intelligence techniques in the development of an enhanced software project management tool, which takes account of the emerging requirement for support systems to address the increasing trend towards distributed multi-platform software development projects. In addressing these aims this research devised a novel architecture and framework for use as the basis of an intelligent assistance system for use by software project managers, in the planning and managing of a software project. This paper also describes the construction of a prototype system to implement this architecture and the results of a series of user trials on this prototype system.1. Software Project ManagementThe key issue in project management is decision making. Software project managers make many decisions every day, ranging from the relatively inconsequential to the significant. Ceteris paribus, good outcomes from those decisions are more desirable than bad outcomes. Project managers make decisions based on a combination of judgement and information from staff, clients, research literature and current market forces, as well as knowledge gained from previous projects. Ideally, all relevant information should be brought together before judgement is exercised. The quality of a decision depends on the adequacy of the available information, the quality of the information, the number of options available at the time of the decision and the ability of the people involved to interpret this information.Software-intensive projects often fail because the project managers lack knowledge of good practices and effective processes which can reduce risk and increase the likelihood of success. Managers of projects need to know how to establish a set of processes which are tailored to a project’s requirements in terms of functionality, time, cost, quality and their associated risks.Constructing, maintaining and extending large complex software systems pose many problems of managing all the people, systems and agencies involved. Although many project management systems are currently available, the enormous scope and complexity of current software systems means moving beyond the current state of practice (for example, PERT charts or Microsoft Project). Managing large-scale projects requires facilities for coordinating independent activities and managing the project plans themselves.To support project managers, organisations have sought to develop tools to assist with various aspects of the management of their software processes. Many tools exist in the market today that assist a project manager in addressing some of these objectives. These tools fall into three main categories:• Project planning - tools which are concerned with the scheduling aspects of planning a project, and pay less attention to organisation and methodological aspects of management.• Process management - which support the framework and rules of management of the project’s process.• Risk analysis - tools used at specific stages during a project to assess risk.However, most of these systems fall short of supporting the project manager in their decision making processes and do not offer assistance in representing knowledge about plans and designs, or provide mechanisms for reasoning about plans and designs in flexible ways. Further aspects to supporting the software project manager which are not addressed by today’s support systems are the distributed and cross-platform nature of systems development.This research is motivated by the assertion that users of existing software project management systems could benefit greatly from the inclusion of intelligent assistance techniques in such tools. In addition, such new support systems should provide for the distributed cross-platform nature of modern client-server development.Therefore the objective of this research was to address this shortfall and provide a support tool which will increase the likelihood of success by helping the project manager whohas to make decisions on these issues. Such a tool will encapsulate expert knowledge and make it available to all users. Some of the potential benefits of this approach as applied to the to decision-making process in the domain of software project management are:• Suggestions are made which help the user balance cost, quality and time in making decisions about the use of project resources.• Knowledge is shared about different lifecycle models and why one or another may be most suitable for the users projects.• Measurements are suggested which will enable the user to see how well the project is reaching greater organisational goals and re-plan the ways to reach these goals, if necessary.Even the most experienced project manager may have difficulty knowing the best planning options, even if the critical input parameters of resources, constraints and requirements are known.2. The Role of an Intelligent AssistantThe notion of an intelligent assistant is not new. Indeed, as far back as 399 BC Socrates claimed to have an intelligent assistant, although not in the strictest sense of course. But Socrates did claim to have a non-human companion, which he called a Daemon. Intelligent and always ready to offer good advice, Socrates daemon could be trusted to act without prompting. Real, hard-coded, linguistic and symbolic links abound between Socrates daemon and today’s notion of an intelligent assistant.A software system designed to act as an intelligent team member (or Daemon) could help in the planning and execution of a project. Such an intelligent project assistant could help to preserve knowledge about tasks, to record the reasons for decisions and retrieve information relevant to new problems. They could function as co-workers, assisting and collaborating with the design or operations teams for complex systems. They could also supply institutional memory. They could recall the rationale of previous decisions and, in times of crisis, explain the methods and reasoning previously used to handle that situation.Significant design projects are typically accomplished by teams. An intelligent project assistant could act as design associate [15]. Designs are almost always redesigned; effective redesign requires an understanding of why previous design choices were made and of how these choices achieved or compromised the desired goals; all are vulnerable to loss of important information from changes in design-team membership.In software development projects in particular, an intelligent project assistant can keep track of specifications, design proposals, and implementations for a software project throughout its life cycle. It can record the design decisions of a constantly changing team and also be a repository of solutions and components for new projects. Reasoning techniques can be used to track the (mis)match between specifications and implementations, while analogy techniques can be used to look for existing specifications, components or implementations that match some new requirement.An intelligent project assistant can additionally be of benefit when training new personnel. For many tasks, on-the-job training is extremely effective, providing the trainee with the chance to make real, on-the-spot decisions and see the consequences. On-the-job training is impossible, however, when a bad decision can be disastrous - as in the management of a large complex software development project. Simulations of the project management process, would enable the development of training systems for such situations [5]. These same simulation capabilities are also important when the cost of assembling large groups of people for training is prohibitive.As part of this research a survey of software project management tool users was conducted to obtain an appreciation of the actual state-of-practice by project managers in relation to tool usage, i.e. what do they actually use these tools for and is this consistent with the tool vendors intended usage [13]. In addition, participants were also asked to consider the aspects of intelligent assistance previously discussed and comment on the possible benefits of incorporating this into a project management support system. All of the project managers considered the notion of an intelligent project assistant as a useful addition to the existing range of features in project management tools. In particular, they supported the notion of a tool which could intelligently manage project knowledge, and capture knowledge and lessons learned about projects into a project knowledge base. Apart from the intelligent assistance aspects of this research, the problems associated with organisations having distributed project teams, coupled with multiple hardware platforms was identified by the project managers surveyed, thus highlighting the need tools to operate in a distributed multi-platform environment.3. Implementing Intelligent AssistanceMany solutions have been proposed to the notion of intelligent assistance over the years. These fall under four major categories, Decision Support Systems, Expert Systems, Expert Critiquing Systems and Blackboard systems. In addition, the concept of Intelligent Agents [16] has recently emerged as a potential fifth category. It is proposed that in the complex domain of software project management, a useful tool to support the project manager in the decision making process is likely to be a hybrid of a number of techniques, including DSS, ES, ECS and the blackboard model. It has therefore been proposed to incorporate the information gathering and analysis techniques of DSS, with the ability of ES to propose possible solutions using expert knowledge and best practices and the power of ECS to critique the possible solutions, thus providing the project manager with every facility to make an informed and quality decision [11] [12]. It is considered that an agent based system will provide for an approach which enables the inter-working of a variety of well understood techniques within a single underlying framework - that of agent-orientated system. We therefore propose a system composed of a library of intelligent software agents - where each agent would play the role of a ‘mini-expert system’ or ‘mini-critiquing system’, each with an associated knowledge base. These agents would utilise the blackboard model of problem solving to converge on possible solution states and examine those states to assess their suitability given current conditions. This agent-orientated system would operate within the overall framework of a decision support system, which would provide for the gathering and analysis of data regarding a project and the development of models of the project with the aid and critique of the agents.The major benefits perceived of this approach are facilitating and improving the quality of decision making by a software project manager by reducing information overload and augmenting the cognitive limitations and bounds of the decision maker. This hybrid method of assistance coupled with the architectural properties of intelligent agents (dynamic and distributed objects) present an ideal strategy to implement intelligent assistance systems in the domain of software project management.In addition to the properties outlined above, an agent-orientated architecture is a natural choice to address the issue of heterogeneous client-server systems development. Recent research in Java-based agents [14] [2] and mobile Java-based agents [9] have concluded that they are a viable technology on which to establish a platform independent agent-orientated architecture. To address the distributed client-server issue, research conducted at San Jose State University [10] successfully used CORBA (Common Object Request Broker Architecture) as a basis for developing platform independent client-server systems, including agent-orientated systems. To provide the necessary flexibility for the proposed system and to tackle the issues above, it is considered that both Java and CORBA provide an appropriate framework on which to base our proposed system.Figure 1 Decision Making ParadigmThe decision making paradigm of this agent-orientated approach is illustrated in Figure 1, where two entities, agents and a user, working together to contribute what they know about the domain to solving some problem and hence make a quality informed decision. The users primary role is to generate and modify solutions; the agents is to analyse those solutions and produce a critique and advise on a possible solution for the human to apply in the next iteration of this process. In this scenario agents constantly ‘watch’ the actions of the user by way of monitoring project parameters as the user developing project plans and inputs actual project data. When an agent has all the information it needs, it will proceed with its analysis and produce its conclusions. These conclusions would be given to the user in terms of advice/criticism on the current/predicted future project situation, and also may be used as input data to other agents. For example, there may be a number of agents who specialise in the selection of the most appropriate process model (lifecycle) for a particular project. The agents could have a set of criteria based on certain attributes of the project such as: the problem domain, product, available resources, personnel, and organisational attributes. These attributes would be examined and for each process model a comparative rating produced, indicating the most appropriate choice of process model. The agent would then communicate these findings to the user and other agents.4. System ArchitectureSoftware architecture focuses on three aspects of software design [1];• Partitioning - The functional partitioning of software modules.• Interfaces - The software interfaces between modules.• Connections - The selection and characteristics of the technology used to implement the interface connections between software modules.This ‘partitioning’ approach was taken to the development of a suitable architecture to support the proposed intelligent assistant system. In this sections we will describe the system architecture from a high level and examine some of the modules and their connections.Figure 2 System ArchitectureA high-level (user level) view of the system architecture is illustrated in Figure 2 and consists of the user interface to the system, the decision support system itself and the underlying knowledge base which contains the expertise and knowledge which will be used to assist the project manager in the planning, managing and execution of a software development project.The component modules of the system architecture are illustrated in Figure 3, and are described below:• User Interface - This component handles the management of all the screen elements (menus, dialog boxes, etc.), validates data entered by the user and passes on clear functional messages to the rest of the system.• System Kernel - This is the core component of the system and handles all the processing and storage of user entered data. It manages all aspects of project plans and channels advice from the agents to the user.• Data Manager - This component manages all aspects of the mapping from the logical view of data to its physical storage and maintenance. This module is under the control of the System Kernel and all requests for data must be channelled through the Kernel.• Agent Controller - This module acts as a controller (or supervisor unit) over the agent community and manages the scheduling and execution of agents, as well as governing write access to the Blackboard.• Blackboard - This represents the global problem solving state of the system. Over time, agents produce changes to the Blackboard which lead incrementally to advice on the project under consideration. The Blackboard is under the control of the Agent Controller and all requests for data read / write must be channelled through the Agent Controller.• Agent Library - All agents are contained in an agent library, but remain under the control of the Agent Controller. The purpose of the Agent Library is to manage the physical agents themselves and to service requests for agent interactions from the Agent Controller.Figure 3 Component ArchitectureCORBA [3] was chosen as the interface bus on which to implement message passing between each of the modules in the system architecture. The CORBA bus allows transparent access to distributed objects over aheterogeneous network of machines and operating systems. CORBA distributes messages via its Object Request Broker (ORB) transparently between registered objects. The ORB receives requests from a ‘client’ to send a message to an object. The broker locates the object referred to by the client and delivers the message to that object. This style of architecture combined with the flexibility of CORBA provides a unique solution to the requirements of independence of implementation language, capacity for evolution and interfacing ability.The use of CORBA allows us to maintain the tool kernel on a typical server while porting the GUI to a client machine, with both the agent library and the project database located physically anywhere on the network. With this in mind, possible alternatives would be:• A ‘thin client’ - possibly web based or implemented in Java, which would therefore be platform independent, while still providing a multimedia oriented GUI.• Classical server-side application containing the kernel, agent supervisor and database.• An agent library located anywhere on the LAN or an intranet/internet, which may be implemented in any language and accessed via CORBA.The main advantages of this approach are:• The ‘light-weight’ clients are platform independent, thrifty in resources, and easily upgradable.• The application has a powerful classical kernel, while retaining the advantages of client-server computing.• It provides a facility for the tool to evolve new services, which can be added as new modules.• The use of a CORBA bus provides interfacing capabilities with other (future) CORBA complaint Various research projects have investigated languages for implementing intelligent agents in recent years [16] [17]. In the early stages, agents were local to individual projects and their languages were mostly idiosyncratic. As a result there are a large number of representation languages, each with their own particular characteristics, which do not have inter-agent communication capabilities. An obvious solution is to have a lingua franca, where ideally all agents that implement the same lingua franca would be mutually intelligible [8]. However, the agent community is still a long way from attaining this goal.An example of a popular intelligent agent language is JESS (Java Expert System Shell) [14] which is a clone of the core of the CLIPS (C Language Integrated Production System) [4] expert system shell developed. JESS contains the main features of CLIPS and is downward compatible with CLIPS, in that every valid JESS script is a valid CLIPS script. The primary representation methodology in both CLIPS and JESS is a forward chaining production rule language based on the Rete algorithm.Currently, only a small number of Java based products for the development of expert system tools exist in the market place. For example, [6] [7] reviews five commercially available tools: Advisor/J (Neuron Data Inc.), Ilog Rules for Java (Ilog Inc.), CruXpert (Crux Inc.), Selectica SRx Selection Engines (Selectica Inc.) and JESS (Sandia National Laboratories). The JESS system was chosen as the primary knowledge representation system for a number of reasons, with the primary motivation that (from an architectural perspective) it is based on an open, mature and portable knowledge representation language, i.e. CLIPS.5. Prototype ImplementationThe development platform chosen was a standard Intel Pentium PC running the Windows NT 4 operating system connected via a LAN to several Intel Pentium PC servers and Sun Microsystem servers and workstations running the Solaris operating system. For the development of this system, two main tools were used: a Java compiler - in this case Sun Microsystems JDK - and a Java implementation of a CORBA ORB - in this case Iona’s OrbixWeb.An early prototype of the proposed system was developed for a number of reasons: Firstly as a proof of concept - the prototype assisted in highlighting any flaws in the proposed architecture and investigated the feasibility of the distributed architecture based on CORBA and Java. It also assisted in identifying possible communication, data storage and knowledge representation problems that may occur.The prototype consisted of a number of servers in the agent architecture - Agent Controller, GUI, and the Agents themselves. The advantage here is that each of these servers are asynchronous from each other and from the clients that call them, making them almost completely independent. Thus on the majority of occasions when the client call is made it can continue with its own process, the call is one-way and so does not have to wait for the calls completion.This initial prototype system was successfully developed and deployed in a distributed manner over a number of networked machines. For example, the expert agent server was run on a Windows NT server, with the client (GUI) deployed on Windows 95, with agent communication taking place (via CORBA) over the network using TCP/IP.Following from this initial prototype a series of four further prototype systems were constructed over an extended period of time, with a phased approach to the evolution of system services being adopted. The final prototype consisted of a set of 30 expert advisory agents in the areas of project planning/re-planning, risk management and metrics and a full set of basic services including the ability to produce multiple scenarios based on the ‘current’ status of the project. As the end of each prototype development stage the current system was demonstrated to a representative group of software project managers as outlined in section 6. Currently the prototype system is undergoing a commercialisation phase, with a view to launching a product in the tool market place.6. Trial UsageThere are several reasons for conducting user trials of the prototype system; Firstly it exposes the system to ‘real world’ project managers and obtains feedback from them. In addition it provides a mechanism to elicit opinion from users as to the added value of the system as compared to traditional project management systems. The trials were conducted using twelve project management staff from two organisations. These staff ranged in experience from novice to highly experienced, with the projects under their control ranging from small scale to large, highly complex systems. The actual user trials were conducted in four distinct stages, one for each of the major prototype releases (not including the first prototype, as it was an architectural proof):• Trial 1 - was conducted using the second prototype and had as its purpose the aim of testing the user interface with respect to data capture.• Trial 2 - was conducted using the third prototype and had the dual purpose of the testing of scenarios and generated advice associated with them.• Trial 3 - was mostly concerned with testing the total functionality of the system.• Trial 4 - was conducted using the final prototype. This trial concentrated on advice produced by the agents.The main output of these trials was a set of four review documents - one for each trial - which detailed the comments and opinions of the users involved in each trial. The following are some of the main finding of these user trials:• Operation - The prototype system was successfully operated by a number of users on a variety of machines. This was an indication that the prototype system was capable of being executed in a commercial environment, although the slow speed of executionwas an important issue. However, users acknowledged that the speed issue was not of great importance for a research prototype, but would be for a commercial version.• Decision support - The general feeling of users was that the prototype system demonstrated that the notion of intelligent assistance for software project management was feasible. In addition, they considered that the prototype implementation provided a suitable framework for supporting decision making and had the potential to be of use in a commercial setting.• Project descriptions - The general opinion of users was that the mechanisms of describing projects (via models and scenarios based on a model) was an appropriate and useful device to capture information about a project. In particular the notion of multiple scenarios to examine multiple views (with corresponding advice) of a project was useful.• Advice - Of paramount interest in these trials was user feedback in relation to advice produced by agents. The overall trend was that novice users considered the advice appropriate and useful as either a reminder of a particular aspect of management, or as an indicator of which direction to consider. However, more experienced project managers expressed the desire for more specific and quantitative advice.• Training tool - A suggestion put forward by a number of users was the possibility of a repositioning of the system for use as a training tool, in which users could develop a model of a fictitious project and thus practice project management skills on a ‘virtual project’.The most difficult issue to tackle which arose during the user trials was the request for advice which was more quantitative in nature. This had proved difficult for two reasons; Firstly, little suitable source material was available which contained quantitative data / results that could be used as the basis for agents. Secondly, it is difficult for humans to discern the differences between quantitative values at a fine grain level with domains such as software project management. For example, there is no appreciable difference between the values of 70% and 75% if they were expressed as a measure of suitability for a given lifecycle model. However, it is worth noting that this quantitative issue - while an important issue in its own right - is not a central issue to the proposed architecture of this thesis. It is however an indicator of the nature of advice users perceive to be useful in addition to the advice already produced. The comments received from users were based on a series of research prototypes and indicate the proposal of an intelligent assistant system for software project management is a viable notion.7. ConclusionsThis paper has set forth a proposal for an intelligent assistant system for use by software project managers. Such an intelligent project assistant could help to preserve knowledge about tasks, function as co-workers, assisting and collaborating with the design or operations teams for complex systems.This research reported in this paper has proposed a novel architecture for the development of the above intelligent assistant system. This approach is a fusion of a number of techniques within a multi-agent framework which aims to improve the quality of the decision making process in the less well understood domain of software project management. This framework incorporates the information gathering and analysis techniques of a Decision Support System with the ability of an Expert System to propose possible solutions using expert knowledge and best practices and the power of an Expert Critiquing System to critique the possible solutions, thus providing the project manager with every facility to make an informed and quality decision. This novel approach enables the inter-working of a variety of well understood techniques within a single underlying framework. An important characteristic of this approach is the combination of these techniques in an open distributed environment with the potential for continuous evolution.To assist with validating the proposed architecture, a prototype system was developed as part of this research and a series of trials conducted in a commercial environment using software project managers. The conclusion of these trials was that the prototype system demonstrated that the notion of an intelligent assistant system for software project management was a viable commercial concept. Further, the prototype system demonstrated that the proposed architecture provided a suitable framework for supporting decision making and had the potential to be of use in a commercial setting.One of the significant drawbacks in relation to the evaluation of the system described in this paper - or indeed any software engineering tool - is that a comprehensive evaluation study requires an extended period of time with access to a large group of potential users. However, in this is a luxury not afforded to most academic research projects. Notwithstanding the foregoing, it is considered that the research reported in this paper provides a significant step forward in the development of a new generation of intelligent assistant systems for software project management.References[1] W.Brown, R.Malveau, H.McCormick and T.Mowbray,“Anti Patterns - Refactoring Software Architectures and Projects in Crisis”, Wiley, 1999.[2] A.Caglayan and C.Harrison, “Agent Sourcebook”, Wiley, 1997.[3] “CORBA: Architecture and Specification”, Object Management Group, 1996.[4] J.Giarratano and G.Riley, “Expert Systems - Principles and Programming”, PWS Publishing Company, 1994. [5] B.Grosz and R.Davis (Eds.), “A Report to APRA on Twenty-First Century Intelligent Systems”, American Association for Artificial Intelligence, 1994.[6] C.Hall (Ed.), “Intelligent Software Strategies”, Cutter Information Corp., Summer 1997.[7] C.Hall (Ed.), “Intelligent Software Strategies”, Cutter Information Corp., Fall 1997.[8] M.Huhns and M.Sing, “Conversational Agents”, IEEE Internet Computing, Vol. 1, No. 2, 1997.[9] nge and M.Oshima, “Programming Mobile Agents in Java - with the Java Aglet API”, technical Report, IBM Research, Japan, 1997.[10] R.Orfali and D.Harkey, “Client/Server Programming with Java and CORBA”, Wiley, 1997.[11] R.O'Connor, T.Renault, C.Floch, T.Moynihan andbelles, “Prompter - A Decision Support Tool using Distributed Intelligent Agents”, In Proceedings of EXPERSYS-97, 1997.[12] R.O’Connor and T.Renault, “Designing an Internet Enabled Decision Support Tool in the Domain of Software Project Management”, In Proceedings of EIS-99, 1999. [13] R.O’Connor and J.O.Jenkins, “Supporting Effective Software Project Management and Control by the use of Intelligent Knowledge-based Guidance”, In Proceedings of 9th European Software Control and Metrics conference (ESCOM), pp. 143 - 151, Rome, Italy, 1999[14] M.Watson, “Intelligent Java Applications”, Morgan Kaufmann, 1997.[15] D.Weld (Ed.), “The Role of Intelligent Systems in the National Information Infrastructure”, AI Magazine, Fall 1995.[16] M.Wooldridge and N.Jennings, “Intelligent Agents: Theory and Practice”, Knowledge Engineering Review, Vol. 11, No. 2, 1995.[17] M.Wooldridge, J.Muller, M.Tambe (Eds.),“Intelligent Agents II: Agents Theories, Architectures and Languages”, Lecture Notes in Computer Science 1137, Springer Verlag, 1995.。