Artificial General Intelligence Technology For Automated Life-Cycle .

1y ago
15 Views
2 Downloads
1.51 MB
16 Pages
Last View : 6d ago
Last Download : 2m ago
Upload by : Jacoby Zeller
Transcription

Artificial General Intelligence Technology forAutomated Life-Cycle Knowledge Management*Edward Chow1, Zack Nolan1, Tony Barrett1, Mark James1, George Paloulian1,David Horres1, Farrokh Vatan1, and Gilbert Torres21JetPropulsion Laboratory, California Institute of Technology2Naval Air Warfare Center, Point MuguMay 11, 2017* The research described in this paper was carried out at theJet Propulsion Laboratory, California Institute of Technologyand was supported in part by the U.S. Department of Defense,Test Resource Management Center, Test & Evaluation/Science and Technology(T&E/S&T) Program under NASA prime contract NAS7-03001, Task Plan Number 81-12346.

AUDREY: Next Generation Artificial IntelligenceVision: Scientist in a Box in a BoxLess Artificial, More Intelligent Machines with human ingenuity More than information: insight Foster new era of autonomyJPL and Beyond Opportunistic data analysis Long-endurance UAV sensing Scientist in a Box Europa missionHey, Rover! Thatrock looks cool.Wanna take a look?Sure thing,AUDREY!2

State-of-the-Art AI Technologies and ChallengesModel Based, Deep Learning, or Human Expertise Traditional Artificial Intelligence (AI) difficultywith uncertainty- Rule-based Artificial Intelligence (AI)approaches fail in real-world problems withmissing and contradictory data Deep Neural Network (DNN) is better withimage, speech, and text but- Requires large training set which is costly inT&E environment and complete retraining oneven slight changes to problem domain- Does not understand big picture meaning likehuman- Problem with real-time testing due to separatetraining and reasoning phases in DNN Machine Learning techniques require expertsto do feature engineering- Need expensive Ph.D. level data scientists forT&E

AUDREY (Assistant for Understanding Data throughReasoning, Extraction, & sYnthesis) AUDREY use bio-inspired Neural Symbolic Processing-Mixed neural and symbolic processing by achieving neural processing atsymbolic level for higher level cognitive reasoning AUDREY leverage human intelligence to achieve better machine intelligence AUDREY capabilities:-Reasoning and learning new knowledge at the same timeDeal with missing or contradictory dataAutomatically synthesize workflows to answer questionsLearn from human and a community of Audrey nodesThe Evolution of AIAchieves unprecedented levels ofreasoning for previously unsolvableproblems

Audrey Architecture Approach Architect Audrey based on a combination ofNeuroscience and Cognitive Science brain-processes-information/5

High Level Cognitive Reasoning withMathematical Logic Grounded in Mathematics– Non-Axiomatic Logic (NAL) Knowledge– Concepts are understood in termsof relationships Truth– The “frequency” and “confidence”of a relation respectively quantifyits truth and evidential support Inference– NAL creates new knowledge bydeductive & non-deductiveinference; however, deductionyields more confident conclusions– NAL can handle summarize similarsemantic statements, even if theyare contradictory, by merging thestatements into a summary, whichis more “confident” than eitherindividual statement.SQPConclusion: “S and P can be summarized by Q”Revision6

Memory Networkgull[1.00, 0.90]swimmerrobin[1.00, 0.90]feathered creature[1.00, 0.90][1.00, 0.90]crowbirdswan7

Audrey Thinking, Learning, and Action ProcessTool LibraryHuman-Like Reasoning,Learning and PlanningPerceptionMulti-levelMemory MatchingDeepNeuralNetworkContext-basedNAL MemoryMatchingAudrey WorkingMemory (Hypothesisand Plans)Audrey AttentionFocused Human-LikeReasoning and line)Intuitive Modeling ingExternal Data SourcesWorkflow Synthesizer(Voice, Video, and Data)8

CNN Examples

Audrey Technologies Audrey has around 1.5 million line of Government ofthe Shelf (GOTS) code Audrey backend system (CORTEX) uses open sourcesoftware code to enable scalable cloud operationsCORTEXAudrey Software Components:– Hunter - natural languageunderstanding system– Kings - Knowledge GraphSynthesizer– Awesome - AutomatedWorkflow Synthesizer– HLR – Human-Like Reasoner– HSG – Hypothetical ScenarioGenerator10

CORTEXCategoricalObjectRepository forTheories andEXperiences NeuronEach componentis built anddeployed into aDockercontainer. So each one isprimed to run ona cloudenvironment.1

Audrey System ArchitectureMonitoringDatabasesSIGINTUAVsExisting andEmerging Real-timeSocialData SourcesMediaTest & EvaluationHardware inthe LoopData Source Pub-Sub DistributionLibrary of Analytic FunctionsWorkflowExecution(KNIME, Hadoop,MapReduce)High-Speed Tool Interface BusQuestions/Responses andData Visualization of ResultsInteractions with userHypotheticalScenario GenerationUnstructuredData ExtractionHuman-likeReasonerTest DataExtracted DomainKnowledgeExternal Knowledge SourcesInteractions withAudrey communityand collaborativehuman expertsNatural LanguageInterfaceWorkflow SynthesisPlug-n-Play ArchitectureClient Interface BusLearnedKnowledge12

TRMC T&E S&T C4T Funded RAID ProjectRAID (Real-time Automated Insight Engine for Data to Decision) RAID will develop an Intelligent, automated assistant for data to decision for F-35-Learn T&E know-how, experiences, and relationships from testers and analystsAssist human in processing large amount of test data in complex situationsUse data to empirically validate and improve learned knowledge with human assistanceUse human-like reasoning to identify insights from structured and unstructured dataEnable distributed testers to use shared knowledge to identify critical test Front-End Data/ImageProcessing and MachineLearningIntelligent InsightExtraction, WorkflowSynthesis, and KnowledgeManagement* DART: Data Acquisition, Recording and TelemetryQueries, Rules, DirectionsDisplay, Advice, Alert, andKnowledgeRAIDRAID use learned knowledge to assist testers to turn data into decision13

RAID Phase 1 Results Demonstrated RAID can learn human T&E knowledge from naturallanguage tidbits Demonstrated RAID can automatically reuse human primed T&Eknowledge to assemble workflow to perform automated ElectronicSupport Measures (ESM) T&E matching operations Demonstrated RAID can learn from data to improve human primedknowledgeRAID* Human-likeProcessing ToAssist Analystto AnalyzeMassiveAmount of DataKnowledgeHumanAssisted AI* Help Humanto ImproveKnowledgeDT, OTAudrey* Life-cycleKnowledgeManagement14

RAID Phase 1 Results Achieved more than 2 orders of magnitude improvement in discerningambiguous ESM matches through the Audrey knowledge learningprocess. Demonstrated automated “unknown-unknown” insight discovery. Demonstrated the model-free Audrey learning system can outperformhuman developed DIVA tool in ESM matching performance.TidbitsRAIDNew PossibilityDevelop more software toolsLearnScalable and Flexible Tools usingHuman Primed and Human AssistedModel-free learning system15

Summary Audrey is a revolutionary Artificial GeneralIntelligence software TRMC T&E S&T C4T RAID technologies enableautomated F-35 life cycle knowledgemanagement16

Automated Life-Cycle Knowledge Management* Edward Chow. 1, Zack Nolan , Tony Barrett. 1, Mark James , George Paloulian , David Horres. 1, Farrokh Vatan , and Gilbert Torres. 2. 1. Jet Propulsion Laboratory, California Institute of Technology. 2. Naval Air Warfare Center, Point Mugu. May 11, 2017 * The research described in this paper was .

Related Documents:

Bruksanvisning för bilstereo . Bruksanvisning for bilstereo . Instrukcja obsługi samochodowego odtwarzacza stereo . Operating Instructions for Car Stereo . 610-104 . SV . Bruksanvisning i original

Artificial Intelligence -a brief introduction Project Management and Artificial Intelligence -Beyond human imagination! November 2018 7 Artificial Intelligence Applications Artificial Intelligence is the ability of a system to perform tasks through intelligent deduction, when provided with an abstract set of information.

10 tips och tricks för att lyckas med ert sap-projekt 20 SAPSANYTT 2/2015 De flesta projektledare känner säkert till Cobb’s paradox. Martin Cobb verkade som CIO för sekretariatet för Treasury Board of Canada 1995 då han ställde frågan

service i Norge och Finland drivs inom ramen för ett enskilt företag (NRK. 1 och Yleisradio), fin ns det i Sverige tre: Ett för tv (Sveriges Television , SVT ), ett för radio (Sveriges Radio , SR ) och ett för utbildnings program (Sveriges Utbildningsradio, UR, vilket till följd av sin begränsade storlek inte återfinns bland de 25 största

Hotell För hotell anges de tre klasserna A/B, C och D. Det betyder att den "normala" standarden C är acceptabel men att motiven för en högre standard är starka. Ljudklass C motsvarar de tidigare normkraven för hotell, ljudklass A/B motsvarar kraven för moderna hotell med hög standard och ljudklass D kan användas vid

LÄS NOGGRANT FÖLJANDE VILLKOR FÖR APPLE DEVELOPER PROGRAM LICENCE . Apple Developer Program License Agreement Syfte Du vill använda Apple-mjukvara (enligt definitionen nedan) för att utveckla en eller flera Applikationer (enligt definitionen nedan) för Apple-märkta produkter. . Applikationer som utvecklas för iOS-produkter, Apple .

BCS Foundation Certificate in Artificial Intelligence V1.1 Oct 2020 Syllabus Learning Objectives 1. Ethical and Sustainable Human and Artificial Intelligence (20%) Candidates will be able to: 1.1. Recall the general definition of Human and Artificial Intelligence (AI). 1.1.1. Describe the concept of intelligent agents. 1.1.2. Describe a modern .

and artificial intelligence expert, joined Ernst & Young as the person in charge of its global innovative artificial intelligence team. In recent years, many countries have been competing to carry out research and application of artificial intelli-gence, and the call for he use of artificial