Proof Markup Language

From Semantic Portal Wiki

Contents 1 Overview 2 Data Model 3 Vocabulary 4 FAQ 4.1 Q: Where are the authoritative information sources about PML? 4.2 Q: What kinds of applications have used PML? 4.3 Q: How is PML specified? 4.4 Q: What tools are available to create PML? 4.5 Q: What tools are available to view and navigate PML? 4.6 Q: Are there sample files? 4.7 Q: What structures in a PML are document are interesting?

Overview

Proof Markup Language (PML) is a Semantic Web based representation for exchanging explanations including

• provenance information - annotating the sources of knowledge
• justification information - annotating the steps for deriving the conclusions or executing workflows
• trust information - annotating trustworthiness assertions about knowledge and sources

The current version is PML v2.0 (PML2, 2006.06, current version) [1]. For more information, please read Proof Markup Language (PML2) Primer.

references{{#vardefine

pagename|pml 2: a modular explanation interlingua }}

1. [[]]

Data Model

pml-provenance

• Information: A piece of information, e.g. a formula in logic languages, and an utterance/word/phrase/sentence/paragraph/article in natural language.
• Source: the source of information. It is the place where we obtain information
  • Agent: actionalble entities.
    • Inference Engine: an system that is able to produce a justification for a conclusion
  • Document: A physical information container and it is not actionable
• SourceUsage: A usage of the source. It records an access of a source at a certain date time.
• Inference Rule: Inference rules are used to derive conclusions from premises.
• Language: The language used to encode the raw string, e.g. English, Knowledge Interchange Format (KIF) , and N3.
• Format: A plan for the organization and arrangement of specified information. Examples: binary, text, pdf, etc.

pml-justification

• NodeSet: A node set hosts a conclusion and a set of alternative inference steps each of which can justify the conclusion. The term ``node set is chosen because a node set captures a set of nodes (with inference steps) from one or many proof trees deriving the same conclusion.
• InferenceStep: An inference step represents a justification for the conclusion of the corresponding node set.
• Query: A Query is a formal representation of user's question. For example, the content of the query can be '(type TonysSpecialty ?x)' which is encoded in KIF.
• Question: A Question refers to natural language version of a user's query. For example, a question can be "What is the type of Tony's Specialty?"

pml-trust

• FloatBelief: An agent's overall belief on a piece of information quantified by a float value between 0 and 1.
• FloatTrust: An agent's overall trust on an information source quantified by a float value between 0 and 1.

Vocabulary

PML2 ontology has three modules:

• PML Data Structure OWL Ontology http://inference-web.org/2.0/ds.owl
• PML Provenance OWL Ontology http://inference-web.org/2.0/pml-provenance.owl
• PML Justification OWL Ontology http://inference-web.org/2.0/pml-justification.owl
• PML Trust OWL Ontology http://inference-web.org/2.0/pml-trust.owl

FAQ

(Note, these are incomplete notes by a new student. Things are probably wrong.)

Q: Where are the authoritative information sources about PML?

• http://inference-web.org
• http://inference-web.org/wiki/Publications

Q: What kinds of applications have used PML?

(intent is to list these from least mature (samples) to more mature (full applications)) list on inference-web.org

• Tony's specialty (this is the example in the PML Primer)
• PML applied to TPTP (Thousands of Problems for Theorem Provers)
• Finding Ramazi's office
• Stephan is working on the MDSA project, which looks at the sun (?) PML is at http://escience.rpi.edu/workflow/0.1/
• UIMALogic's Mississippi Automated System (a text extraction application)
• UTEP's gravity sets

Q: How is PML specified?

PML is specified using three ontology modules:

• PML Provenance (PML-P),
• PML Justification (PML-J), and
• PML Trust (PML-T).

PML-J imports PML-P.

PML-T imports PML-J (and PML-P by transitivity).

Of lesser importance is DS, which PML-P imports for some fundamental "data set" descriptions (typed rdf:List).

Q: What tools are available to create PML?

• Cynthia wrote a Java PML API

Q: What tools are available to view and navigate PML?

• IW Browser was developed by Cynthia
• Probe It! was developed by UTEP
• POMELo is an early tool by Alvaro

Q: Are there sample files?

Sample PML files are listed at http://inference-web.org/wiki/PML_Datasets

Q: What structures in a PML are document are interesting?

SPARQL queries for Proof Markup Language instance data