120 likes | 246 Views
Strategic Research Directions in AI: Knowledge Representation, Discovery, and Integration Intelligent Information Systems Institute IISI Organizing Committee: Carla Gomes David Lewis Bart Selman In collaboration with AFRL/IF. Goal of Workshop.
E N D
Strategic Research Directions in AI:Knowledge Representation, Discovery, and IntegrationIntelligent Information Systems InstituteIISIOrganizing Committee:Carla GomesDavid LewisBart SelmanIn collaboration with AFRL/IF
Goal of Workshop Provide the AFRL/IF with concrete research recommendations in Knowledge Representation, Discovery, and Integration. Identification of key research directions to pursue in the next decade. Output of workshop: report of recommendations
TECHNICAL- GAPs in the coverage of technology FUNCTIONAL- GAPs in the coverage of requirements from Air Force MAJCOMs Begin with MAJCOM FUNCTIONAL Requirements Begin with total capability for a technology area GAP in Technology Coverage subtract R&D work done at: subtract Requirements Satisfied or N/A to Technology Area AFOSR Other AFRL Directorates Other Service Labs Other DoD Agencies Commercial Sector Colleges/Universities Other Countries AFRL/TDs Total Capability Needs Result ARL & CECOM NRL & SPAWAR NRO NIMA Technology Coverage by External Groups AFRL/IF Investment Strategy Covers the Entire Capability Needs DARPA Result GAP in TECHNOLOGY Coverage International, Academia Commercial Sector Gap Analysis Opportunities for IF S&T Investment
Drill Down ProcessDesired Capabilities to Technical Gaps Capability Goals Capability Objectives Capability Gaps Technical Challenges Technical Required Technical Gaps Note: Researchers only address lower levels (technical challenges)
Capability Goals and Objectives • Goal: Assess Global Conditions and Events • Perform Predictive Battlespace Awareness • Assess Enemy Attacks against Friendly Assets • Assess in Realtime Effect of Friendly Operations • Characterize emerging threats • Goal: Establish and Maintain Battlespace Situation Awareness • Receive/Correlate Information then Fuse/Disseminate/Display in Tailorable Presentation • Monitor World Events, Physical Environs, and National Policy Guidance • Goal: Locate, Identify, Track, Observe, Monitor All Forces Anywhere/Anytime in Near Realtime • Collect against Multiple Targets (Across EM Spectrum, Moving or Stationary, Real or Decoy) despite Opposition • Fuse Data and Information from Various Sensors and Sources • Collect Information in All Operational Environments (Physical and Infosphere) • Goal: Deploy and Employ Independent C3ISR Elements to Forward, Distributed Locations • Provide Battlespace Situation Awareness • Plan and Prepare to Execute Operations
Capability Gaps • Ability to acquire accurate perception • Ability to capture knowledge of past history • Ability to comprehending existing situations • Ability to transfer the situational picture • Control Optimization • Intent : Continuous real time collection management during execution (plan refinement, reaction to enemy actions, taking advantage of opportunities) • Ability to measure the quality of decision level information
Technical Challenges • Acquire accurate perception (i.e. data) • Collecting all-source data • Structured Representation • Semantic Mediation of heterogeneous sources • Comprehending existing situations • Rapid Knowledge Formation • Knowledge Discovery • Identify new and unknown links and relationships • Prediction • Captured knowledge of past history • Model Generation • Human assisted Model generation • Automatic Model Generation from past experience • Data/knowledge base tools • Knowledge Engineering/Acquisition • Large Scale Virtual Repositories • A priori Knowledgebase for Fusion • Data Mining
Workshop focus areas:Technical ChallengesinKnowlegde Representation, Discovery, and Integration
Example of template for panel members – “hot” topics • Title of topic: • Objective: • Key Challenges:
Example of Template Technical Challenges SUMMARY OF THE STATE OF THE ART Research Areas CURRENT LIMITATIONS
Agenda Thursday June 26 Intro : 900-920 Carla Gomes Craig Anken – AFRL/IF Panel 1: 920-1020 Data & Information Panel – Chair: David Lewis Break: 1020-1050 Panel 2: 1050-1150 Information & Knowledge Panel - Chair - David Jensen Lunch: 1150-1250 Panel 3: 1250-150 Knowledge & Understanding Panel I - Chair - Bart Selman break : 150-220 Panel 4: 220-320 Knowledge & Understanding Panel II (Level 3/4): Chair: Marie desJardins
Agenda (cont.) Thursday June 26 Breakout Sessions: 320-445 Wrap up: 445-530 Dinner: 6:00 Friday June 27 Report Preparation Breakout Sessions: 900-10:30 Break: 10:30-11:00 Wrap-up: 11:00-12:00