IW Meeting 2010-08-12

From Inference Web

Jump to: navigation, search

Contents

Meeting Info

Attendees

  • Cynthia Chang
  • Li Ding (away)
  • Tim Lebo
  • Paulo Pinheiro da Silva (away)
  • Deborah McGuinness (away)

IW Overall Objectives

     1) PML ontology: (OWL level only) potential non-monotonic  changes (Li) 
     2) PML practice  
         2.1) Use cases and examples (Tim) 
         2.2) RDF level issues (Cynthia) 
         2.3) PML Tools / API issues (Cynthia) 
         2.4) PML primer (Tim) 
     3) PML and OPM mapping (Paulo) 
     4) IW management (TBD)
        4.1) inference web infrastructure and research plan
        4.2) technologies for engagement 

Today's Agenda

  • Incubator Group
    • 3 scenarios, first is done.
    • references are only useful if cited by the scenario. (tag in medelay and add to wiki page)
    • they need people to describe how the reference relates
  • Two distinct NodeSets isConsequentOf the same InferenceStep?
  • How will the current PML tools be affected by naming Information with the value of pmlp:hasURL and omitting pmlp:hasURL?
  • pmlp:isConsequentOf [ a pmlp:SourceUsage ]
  • accept bapsm2309 to the iw google group?
  • PML API documentation
    • difference between library, source, and distribution?
  • Google Code Issues.

Notes from Today's Discussions

PML API examples

Each example in the PML Primer references the java code that produces the PML. http://inference-web.org/wiki/iw:PML_Primer use the currently published one at:http://inference-web.org/2007/primer/


Difference between library, source, and distribution

  • library is just the jar, can code against it (put in CLASSPATH)
  • source expands into org/ - v2 is the JAVA API that supports the ontology 2.0.
  • distrubition is a superset of library and source (grab that and you have everything.)
    • library
    • source
    • primer
    • javadoc

bapsm2309

Tim and Cynthia don't recognizes it.

Two distinct NodeSets isConsequentOf the same InferenceStep?

Can two nodesets refer to the same inference step?

NodeSet has a property called isConsequentOf with range of InferenceStep.

The following example has 2 DISTINCT NodeSets referencing the SAME InferenceStep.

provenance:nodeset_1 a pmlj:NodeSet ;

  pmlj:hasConclusion <http://logd.tw.rpi.edu/source/data-gov/file/1008/version/2010-Jul-21/conversion/data-gov-1008-2010-Jul-21.e1> ;
  pmlj:isConsequentOf provenance:infstep_1 .     <----------------- Is it OK for two NodeSets to point to the SAME InferenceStep? (yes)

provenance:nodeset_2 a pmlj:NodeSet ;

  pmlj:hasConclusion <http://logd.tw.rpi.edu/source/data-gov/file/1008/version/2010-Jul-21/conversion/data-gov-1008-2010-Jul-21.sameas> ;
  pmlj:isConsequentOf provenance:infstep_1 .     <----------------- Is it OK for two NodeSets to point to the SAME InferenceStep? (yes)

provenance:infstep_1 a pmlj:InferenceStep ;

  pmlj:hasInferenceEngine tw_converter:csv2rdf4lod_56eddd08d634faa3dbba4138117db1a8 ;      #### (URI, not bNode)
  pmlj:hasAntecedentList provenance:list_1 .

@prefix tw_service: <http://logd.tw.rpi.edu/source/tw-rpi-edu/service/> .

tw_converter:csv2rdf4lod_56eddd08d634faa3dbba4138117db1a8

 a tw_service:csv2rdf4lod , pmlp:InferenceEngine ;   
  dcterms:identifier "csv2rdf4lod_56eddd08d634faa3dbba4138117db1a8" ;
  dcterms:version "2010-Aug-09" .
  

Cynthia doesn't see anything wrong with the above example.

InferenceStep has property fromAnswer which is the final conclusion of the proof. Is this InferenceStep instance used to conclude one final conclusion(NodeSet)? From http://inference-web.org/2.0/pml-justification.owl

[]

a :InferenceStep;
 pmlj:fromAnswer []
 .

pmlj:fromAnswer points to the root NodeSet, allowing to get to the conclusion without going through the intermediate NodeSets. But in this case, the inference step would be pointing to two DISTINCT NodeSets. An InferenceStep can only have max 1 fromAnswer, which would prove the DISTINCT NodeSets to be the same.

This is one issue bothering me for a while, with combined tptp proofs in particular.


Google Code Issue Form

Summary: pmlj:fromAnswer max cardinality 1

Description: max 1 cardinality prevents an InferenceStep from pointing to more than NodeStep.

CSV conversion case:

Two files are the result of a single invocation of a converter. Each file needs to be justified by the invocation.

TPTP proofs case:

Current Solution: Make a duplicate InferenceStep that has the exact same attributes.



changes: hasURL and pmlp:isConsequentOf [ a pmlp:SourceUsage ]

Cynthia: changes should not take too long.

1) change ontology 2) change API 3) change apps

How do you want to use it?

pmlp:isConsequentOf [ a pmlp:SourceUsage ] would be a leaf, but the nodeset could be annotated with an additional

Cynthia; we need everyone to agree and Deborah to OK it. Cynthia just needs the go-ahead.


Google Code Issues

Potential discussions

Tabled discussions

Actions

  • Tim to verify that URL Lost issue is resolved. and close if OK.
  • DONE: Li to get Tim Google Code privs to close issues.

Summary

Personal tools
Navigation