IW Meeting 2009-04-23

From Inference Web

• 2009-04-23
• IW Meeting 2009-04-16

agenda

• cynthia's issues
  • action-2009-001
• inference-web.org
  • switch to wiki
  • .htaccess change (bug fix)
  • watch dog on onto.rpi.edu
• export pml from wiki,
  • Li Ding, RDF version special:ExportRDF/Li_DING -- Li is a pmlp:Person and his pmlp:name is "Li Ding"
  • download the patched version: http://code.google.com/p/smwbp/source/browse/#svn/trunk/release/semediawiki-1.4.2
• Combine_proof_example_2 discussion
• complexity analysis on traverse algorithm
• proof quality metric discussion
• funding discussion

notes

• PML API - cynthia got a new release, Li want it versioned

The latest version of PML API is available for download:
1. PML library - http://inference-web.org/downloads/pml.jar
2. PML source - http://inference-web.org/downloads/pml_source.tar
3. PML distribution - http://inference-web.org/downloads/pml_distribution.tar.gz

In this version,  the property http://inference-web.org/2.0/pml-provenance.owl#hasConfidenceValue is included.

• iwsearch - no progress
• spcdissearch - no progress
• open provenance model reading - honglei, cynthia read it
  • http://twiki.ipaw.info/bin/view/Challenge/TetherlessPC3 (james did it)

• inference-web.org
  • it is semantic wiki
  • create account should be restricted from public users. any registered user can create account for other.
  • any group member can publish their inference related projects
  • publication management, we will query all's publication info from owner's sites

• export pml from wiki
  • Li did a live demo on declaring Li Ding is an instance of person.

<swivt:Subject rdf:about="&wiki;Li_Ding">
 <rdfs:label>Li Ding</rdfs:label>
 <swivt:page rdf:resource="&wikiurl;Li_Ding"/>
 <rdfs:isDefinedBy rdf:resource="&wikiurl;Special:ExportRDF/Li_Ding"/>
 <rdf:type rdf:resource="http://inference-web.org/2.0/pml-provenance.owl#Person"/>
 <pmlp:hasName rdf:datatype="http://www.w3.org/2001/XMLSchema#string">Li Ding</pmlp:hasName>
</swivt:Subject>

• Combine_proof_example_2 we know the problem better
  • there are altogether four different proofs
  • paulo believe a justification does not need to be as complete as a proof.

 Possible "proofs" for the conclusion of D1: (D1), (D1,C1), (D1,C1,B1),...,(D1,C2),(D1,C2,B2),...

todo

• publication management on inference-web.org (how to used queried results)
• come back: Combine_proof_example_2 backtracking
• come back: complexity analysis on traverse algorithm
  • We now have several equivalent formulations of proof combining: proof combining, proof assembly, And-of-tree. Li and Honglie will work out a set of unified notations this week;
  • Honglei will elaborate a few more examples to Shangguan and Rui so that they can continue to take on the second part of the proof;
• come back: proof quality metric discussion
• come back: funding discussion
• inference web infrastructure:
  • watchdog (cynthia prefer to run it off campus), we should consider VCC support
  • SVN and issue tracker
  • inference web servers? and adding memory to onto.rpi.edu. http://www.crucial.com/store/listparts.aspx?model=PowerEdge%201950
  • version for PML API