The SNePS Research Group

1
The SNePS Research Group
SNeRG
Prof. Stuart C. Shapiro, Prof. William J.
Rapaport Prof. Carl Alphonce, Prof. Josephine
Anstey, Prof. Debra T. Burhans, Prof. Michelle L.
Gregory, Prof. Jean-Pierre A. Koenig, Prof. David
R. Pierce Graduate Students Jonathan Bona,
Trupti Devdas Nayak, Albert Goldfain, Frances L.
Johnson, Michael Kandefer, John F. Santore,
Lunarso Sutanto Undergraduate Students Vikranth
B. Rao, Isidore DingaMadou
Semantic Network Processing System
The long-term goal of The SNePS Research Group is
the design and construction of a
natural-language-using computerized cognitive
agent, and carrying out the research in
artificial intelligence, computational
linguistics, and cognitive science necessary for
that endeavor. The three-part focus of the group
is on knowledge representation, reasoning, and
natural-language understanding and generation.
The group is widely known for its development of
the SNePS knowledge representation/reasoning
system, and Cassie, its computerized cognitive
agent.
Cassie the FEVAHR Agent
GLAIR Architecture
A Plan for Detonating Unexploded Landmines (UXOs)
Grounded Layered Architecture with Integrated
Reasoning
Knowledge Level
all(a)(Agent(a) gt ActPlan(Blowup(a, UXOs),
Act(a, Cascade(SearchforUxo(a),
WithSome(obj, Near(a, obj),
WithNew(ch ex,
Charge(ch), Explosion(ex),
Possess(a, ch),
Cascade(Place(a, ch,
obj), Hide(a),
Waitfor(a, ex), SearchforUxo(a))),
goto(a, SafeZone))))))
NL
SNePS
Perceptuo-Motor Level
Sensory-Actuator Level
Vision
Natural Language Interaction with
FEVAHR (Foveal Extra-Vehicular
Activity Helper-Retriever) Cassie Who are you?
My name is Cassie' and I am the SNePS
cognitive agent. Who did you talk to? I
talked to Stu and I talked to Bill and I talked
to Carl and I talked to David and I talked to
Debbie and I talked to J.P. and I talked to
Josephine and I talked to Michelle and I am
talking to you. Who did you see? I saw Stu
and I saw Albert and I saw Fran and I saw John
and I saw Jon and I saw Lunarso and I saw
Michael and I saw Trupti and I see you.
Sonar
Motion
Proprioception
Example SNePS Ontology
The Trial The Trail is an interactive drama for
an immersive VR environment. Its intelligent
agents are SNePS-driven.
Contextual Vocabulary AcquisitionFrom Algorithm
to Curriculum
Belief Base Revision with Reconsideration

• PIs William J. Rapaport (CSE SNeRG) Michael
  W. Kibby (Learning Instruction Dept.)
• CVA computing a meaning for unknown word from
  contextual clues prior knowledge
• There came a white hart running into the hall
  with a white brachet next to him, and thirty
  couples of black hounds came running after them.
  As the hart went by the sideboard, the white
  brachet bit him. The knight arose, took up the
  brachet and rode away with the brachet. A lady
  came in and cried aloud to King Arthur, Sire,
  the brachet is mine. There was the white brachet
  which bayed at him fast. The hart lay dead a
  brachet was biting on his throat, and other
  hounds came behind. Morte DArthur
• Cassie learns what brachet means From above
  text prior knowledge about harts, animals, King
  Arthur, etc. no info about brachetsInput SNePS
  version of simplified English narrative. Output
  Definition frame (varies with context and prior
  knowledge)
• 1.First Sentence
• A hart runs into King Arthurs hall.
  3. Full Story
• In the story, B12 is a hart. A hart runs
  into King Arthurs hall.
• In the story, B13 is a hall. A white
  brachet is next to the hart.
• In the story, B13 is King Arthurs. The
  brachet bites the harts buttock.
• In the story, B12 runs into B13 The
  knight picks up the brachet.
• A white brachet is next to the hart.
  The knight carries the brachet.
• In the story, B14 is a brachet. The lady
  says that she wants the brachet.
• In the story, B14 has the property white.
  The brachet bays at Sir Tor.
• Therefore, brachets are physical objects.
  prior knowledge only hunting dogs bay
• deduced while reading, using 4.--gt
  (defineNoun brachet)
• prior knowledge only physical objects have
  color Definition of brachet
• 2.--gt (defineNoun brachet) Class
  Inclusions hound, dog,
• Definition of brachet Possible Actions
  bite buttock, bay, hunt,
• Class Inclusions phys obj, Possible
  Properties valuable, small, white,

UAV
Current Belief Base
Time UAV
INTEL TROOPS T1
Red in D1,D2 Red in D1,D3 T2
Red in D1,D2 Red in D1,D3,C3,Bridge-D
T3 Red in D1,D2 Red in
D1,D3,C3,D4 Red in C3 Always TROOPS gt
UAV gt INTEL Always TROOPS gt INTEL gt UAV
INTEL
Asserting beliefs into the belief base (or KB)
Adding them to the KB Stating them to be true.
T1 UAV INTEL disagree on Red troop
location gt contradiction. Consolidation makes a
belief base consistent -- in this case by
removing (or retracting) INTELs statement.
Contracting the KB by INTELs statement. (UAV gt
INTEL) T2 UAV INTEL again disagree. At T3,
BLUE TROOPS confirm an INTEL belief over that of
UAV So, we reverse the INTEL/UAV credibility
order. Thus, UAV is disbelieved. Reconsideration
of the KB is defined as consolidation of all base
beliefs (current, or not). INTELs earlier
beliefs are recaptured ( returned to the KB),
and UAVs are retracted.
BLUE TROOPS
Identifying Perceptually Indistinguishable
Objects Is that the same one you saw before?
Crystal Cassies view of the world showing two
perceptually indistinguishable robots, one of
whom she is following.
Robots used in human subjects and Crystal
Cassies tasks
What Crystal Cassie can see a table with glasses
and a computer lab with two people
SNeRG website www.cse.buffalo.edu/sneps