sami: situational awareness from multi-modal input naveen ashish
Post on 22-Dec-2015
219 views
TRANSCRIPT
SAMI: Situational Awareness from Multi-
modal Input
Naveen Ashish
Talk Organization Why are we at RESCUE interested ? Situational Awareness (SA)
– Introduction System architecture Research challenges Expected outcomes and artifacts Extraction system demonstration
Team
Naveen AshishSharad MehrotraNalini VenkatasubramanianUtz WestermannDmitry KalashnikovStella ChenVibhav GogatePriya GovindarajanRam HariharanJohn HutchinsonYiming MaDawit SeidJay LickfettChris DavisionQuent Cassen
Bhaskar RaoMohan TrivediRajesh HegdeSangho ParkShankar Shivappa
Ron EguchiMike Mio
Jacob Green
Information from Various Sources
People/Victims at disaster
Emergency responders
News, video, audio footage
GIS, satellite imagery, maps
Pushing “Human-as-sensor”
More Data ≠ More Information
SA
Where are the fire personnel ?Have all medical supplies reached ?What areas should we start evacuating first ?
Situational Awareness Wide variety of fields
– Beginning in mid-80s, accelerating thru 90s– Fighter aircraft, ATM, Power plants,
Manufacturing Definitions
– "the perception of elements in the environment along with a comprehension of their meaning and along with a projection of their status in the near future"
– "the combining of new information with existing knowledge in working memory and the development of a composite picture of the situation along with projections of future status and subsequent decisions as to appropriate courses of action to take"
Situational awareness and decision making Areas
– Cognitive science– Information processing– Human factors
Knowing what is going on
Abstraction of Information
Multimodal Input: Text, Audio, Video
Events
Awareness
First-cut Architecture
EVENT BASE
Qu
ery
ing
an
d
A
na
lysi
s
Graph View
VISUALIZATION and USER INTERFACES
SpatialIndexingPDF Histogram
KNOWLEDGE: ONTOLOGIES
Text
Audio
Video
Internet
RAW DATA
EVENT EXTRACTION
REFINEMENTDisambiguationLocation
Centered around EVENTS as fundamental abstractions
Research Areas
Event ModelingEvent Extraction
Disambiguation
Location Uncertainty
Graph Analysis
GIS Querying
Event Modeling What is an event ? Event Representation
RELIABILITY
PEOPLE EVACUATION
LOCATION
TIME
REPORT
TYPE
NAME LOCATION
AGENCY
FROM TO
OPERATION
NUMBER
Domain Knowledge
Captured as Ontologies
ROAD EVACUATION
EVACUATION
AIREVACUATION
IS-A
IS-A
THAILAND
SOUTHERN REGION
…….
PHUKETPHUKET,CHANGWAT
Event Extraction Long history of information extraction
– IR (MUC efforts)– Web data extraction
DARPA ACE– Entities, Relations, Events– Events in 2004
Event extraction accuracy is still low SA Domain
– Stream of information– Duplicated, ambiguous– Reliability– Conversations
Modalities– Text
Semantics Driven Approach
Semantics Driven Challenges
– Framework– Ontologies
What semantics required for event extraction ? Application
With NLP, ML techniques Performance
– SA specific Duplicates, reconciliation, temporal,
conversations …..
Disambiguation
Disambiguation
– point-location
in terms of landmarks
uncertain, not (x,y)
– reasoning on such data support all types of queries
Uncertainty is a Challenge
Report 1: “... a massive accident involving a hazmat truck on I5-N between Sand Canyon and Alton Pkwy ...”
Report 2: “... a strange chemical smell on Rt133 between I405
and Irvine Blvd ...”
Report 1 R
epor
t 2
Implications of Uncertainty in Text How to model uncertainty?
– probabilistic model– P(location | report)
e.g. report says “near building A”
Queries– cannot be answered exactly...
use probabilistic queries all events: P(location R | report)
> 0– SA requirements
triaging capabilities fast response
– top-k – threshold: P(location R | report) >
-RQ, k-RQ, k -RQ
How to map text to probabilities?– use spatial ontologies
AB
R
Graph Analysis
GAAL Inherent spatio-
temporal properties Graphs are powerful
for querying and analysis
Current FGDC Search
GIS Search
Progressive Refinement of Data
GIS Search
Deliverables, Outcomes, Artifacts “Vertical” thrusts
– Event extraction system (TEXT)– Disambiguation system– GIS search system
Overall system demonstration ? “By-products”
– Ontologies Computer science research areas
DatabasesSemantic-Web
Information Retrieval
Intelligent Agents (AI)
Thank you !
http://sami.ics.uci.edu