Enabling Explanations: The Inference Web and PML Approach

1
Enabling ExplanationsThe Inference Web and PML
Approach

• Deborah McGuinness, Paulo Pinheiro da Silva, Li
  Ding
• Knowledge Systems Laboratory
• Stanford University
• info dlm pp _at_ ksl.stanford.edu
• Inference Web and PML is joint work with Fikes,
  Chang, Deshwal, Ding, Narayanan, Glass, Makarios,
  Jenkins, Millar, and has particularly benefited
  from input from Batelle, IBM, SRI, MIT, UMBC, and
  others
• More info http//iw.stanford.edu/documents_pape
  rs.html

2
Outline

• Motivation
• Background - Inference Web
• PML
• Building a PML document by example
• Browsing PML
• Directions and plans
• Discussion

3
Motivation Trust throughTransparency

• If users (humans and agents) are to use, reuse,
  and integrate system answers, they must trust
  them.
• System transparency supports understanding and
  trust.
• Even simple lookup systems benefit from
  providing information about their sources.
• Systems that manipulate information (with sound
  deduction or potentially unsound heuristics)
  benefit from providing information about their
  manipulations.

Goal Provide interoperable infrastructure that
supports explanations of sources, assumptions,
and answers as an enabler for trust.
4
Inference Web Concerns

• Information Manipulation Traces
• hybrid, distributed, portable, shareable,
  combinable encoding of proof fragments supporting
  multiple justifications
• Presentation
• multiple display formats supporting browsing,
  visualization, summaries,
• Abstraction
• understandable summaries
• Interaction
• multi-modal mixed initiative options including
  natural-language and GUI dialogues, adaptive,
  context-sensitive interaction
• Trust
• source and reasoning provenance, automated trust
  inference
• McGuinness Pinheiro da Silva, ISWC 2003, J.
  Journal of Web Semantics 2004

5
Inference Web

• Framework for explaining question answering tasks
  by abstracting, storing, exchanging, combining,
  annotating, filtering, comparing, and rendering
  justifications from question answerers
• IWs Proof Markup Language (PML) is an
  interlingua for proof interchange. Represented
  in OWL
• IWBase is a distributed repository of
  meta-information
• IW Registration and Validation services provide
  support for PML generation, validation, and
  checking
• IW Browser provides display capabilities for PML
  documents
• IW Abstractor provides rewriting capabilities
  enabling more understandable presentations
• IW Explainer provides multi-modal dialogue
  options including alternative strategies for
  presenting explanations, visualizations, and
  summaries
• IW Trust provides methods for propagating trust
  values
• IW Search (enhanced SWOOGLE for PML documents)

6
How to achieve transparency using IW

• Question Answering system gets registered with
  Inference Web (using registration services)
• Question Answering system encodes justifications
  of its answers in PML (using PML generation and
  validator services)
• Application provides access to explanations via
  IWs explainer, browser, and search features

7
One Proof Browser
8
(No Transcript)
9

10
IW Explanation Application Areas

• Theorem proving
• First-Order Theorem Provers Stanford (JTP
  (KIF/OWL/)) SRI (SNARK)
• SATisfiability Solvers University of Trento
  (JSAT)
• Information extraction IBM (UIMA), Stanford
  (TAP)
• Information integration/aggregation USC ISI
  (Prometheus,Mediator -gt Fetch) Rutgers ,
  Stanford (TAP)
• Task processing SRI International (SPARK)
• Service composition Stanford, U. of Toronto,
  UCSF (SDS)
• Semantic matching University of Trento
  (S-MATCH)
• Problem solving University of Fortaleza
  (ExpertCop)
• Trust Networks U. of Trento (IWTrust)
• New collaboration with W3C / MIT
• Govt. Program Usage ARDA NIMD, DARPA PAL,
  SAPIENT, Explainable Knowl. Agg, UIMA
• Company Usage IBM, SRI, Sandpiper,

11
Proof Markup Language

• PML is a proof interlingua
• It can be used to represent justification of
  information manipulation steps done by theorem
  provers, extractors, etc.
• The main components concern inference
  representation and provenance issues such as
  author, source, etc.

12
Structure

13
Querying the Web

• John asks Google 2.0 the following question
• What happened in The Night Club on March 1st,
  1997?
• An Answer There was smoke and fire
• Sources
• The whole place got tons of black smoke.
• http//www.cnn.com/1997/03/03/US/XYZ.html
• fire took hold with devastating speed
• http//news.bbc.co.uk/1/hi/world/279w1119.stm
• Where there's smoke there's fire
• http//www.giga-usa.com/quotes/topics/proverbs_
  t374.htm
• the fire quickly spread, filling the building
  with
• thick, black smoke
• http//www.redcross.org/article/947,00.html

(gtSF)
(SF)
(S)
(F)
(S-gtF)
(SF)
14
Proof Markup LanguageTop-Level Concepts
InferenceStep
SourceUsage
ModelElement
Mapping
ProofElement
ProvenanceElement
NodeSet
Question
Query
15
Proof Markup LanguageNode Sets and Inference
Steps (1/2)
A trivial justification (SF) has been
asserted from the Red Cross (RC) website
Encoding this justification in PML
http//foo.com/Example.owlSmokeFire
iwNodeSet
iwisConsequenceOf
iwInferenceStep
Direct Assertion (DA)
iwhasRule
iwhasSourceUsage
824,1058 on RC
iwhasEngine
CWM
iwhasConclusion
(SF)
iwhasLanguage
N3
16
Proof Markup LanguageNode Sets and Inference
Steps (2/2)
And this is the same NodeSet in XML ltiwNodeSet
rdfabout"http//foo.com/Example.owlSmokeFire"gt
ltiwhasConclusiongt(SF)lt/iwhasConclusiongt
ltiwhasLanguage rdfresource"http//inferenceweb.
stanford.edu/registry/LG/N3.owlN3" /gt
ltiwisConsequentOfgt ltiwInferenceStepgt
ltiwhasIndex rdfdatatype"http//www.w3.or
g/2001/XMLSchemaint"gt0lt/iwhasIndexgt
ltiwhasRule rdfresource"http//inferenceweb.stan
ford.edu/registry/DPR/Told.owlTold"/gt
ltiwhasInferenceEngine rdfresource"http//i
nferenceweb.stanford.edu/registry/IE/CWM.owlCWM"/
gt ltiwhasSourceUsagegt
ltiwSourceUsagegt
ltiwspanFromByte
rdfdatatype"http//www.w3.org/2001/XMLSchemaint
"gt824lt/iwspanFromBytegt
ltiwspanToByte rdfdatatype"http//www.w3
.org/2001/XMLSchemaint"gt1058lt/iwspanToBytegt
ltiwhasSource rdfresource"http//
inferenceweb.stanford.edu/registry/PUB/RC.owlRC"/
gt lt/iwSourceUsagegt
lt/iwhasSourceUsagegt lt/iwInferenceStepgt
lt/iwisConsequentOfgt lt/iwNodeSetgt
17
Justification Collections and Proofs
iwhasAntecedent
Direct Assertion (DA)
AND Intro (I)
from RC
SF
18
The Combine Operation
Combine(A,B) C
B
S
F
Direct Assertion From BBC
Direct Assertion from CNN
C
A
Direct Assertion (DA)
AND Intro (I)
from RC
SF
19
A Result of Combining Justifications
C
20
SWOOP Class View

21
Protege

22
SWOOP Property View

Swoop 2.2.1 loaded with iw.stanford.edu/2004/0
7/iw.owl
23
Proof Elements

24
Provenance Elements

25
Potential Usage in TAMI

• PML as a justification interlingua
• Inference Web tools for browsing, abstracting,
  summarizations, etc.
• IW Registry for components
• IW Search for finding proofs
• Next steps
• Review PML for representational adequacy (chris
  issue of markers for ignoring terms, )
• Register question answerers (CWM, Pychyncho?,
  TMS)
• Augment question answerers to generate PML
• Review output leveraging existing strategies and
  tactics
• Generate new strategies and tactics as needed

26
Discussion

http//iw.stanford.edu/ http//iw.stanford.edu/20
04/07/iw.owl