Abstracting web agent proofs into human-level justifications

From Semantic Portal Wiki

Edit

Reference:

1. Vasco Furtado, Paulo Pinheiro da Silva, Deborah L. McGuinness, Priyendra Deshwal, Dhyanesh Narayanan, Juliana Carvalho, Vladia Pinheiro, Cynthia Chang. Abstracting Web Agent Proofs into Human-Level Justifications , Proceedings of the 20th International FLAIRS Conference (FLAIRS-20) pp.80-85, 2007

bibtex

@inproceedings { KSL-07-06 ,
  author = "Vasco Furtado, Paulo Pinheiro da Silva, Deborah L. McGuinness, Priyendra Deshwal, Dhyanesh Narayanan, Juliana Carvalho, Vladia Pinheiro, Cynthia Chang", 
booktitle = "Proceedings of the 20th International FLAIRS Conference (FLAIRS-20)", 
pages = "80-85", 
title = "Abstracting Web Agent Proofs into Human-Level Justifications", 
year = "2007", 
} 

abstract: Information supporting answer explanations are derived from proofs. One ofthe difficulties for humans to understand web agent proofs is that the proofsare typically described at the machine-level. In this paper, we introducea novel and generic approach for abstracting machine-level portable proofsinto human-level justifications. This abstraction facilitates generatingexplanations from proofs on the web. Our approach consists of creating arepository of proof templates, called abstraction patterns, describing howmachine-level inference rules and axioms in proofs can be replaced by rulesthat are more meaningful for humans. Intermediate results supportingmachine-level proofs may also be dropped during the abstraction process.The Inference Web Abstractor algorithm has been developed with the goal ofmatching the abstraction patterns in the repository against the originalproof and applying a set of strategies to abstract the proof therebysimplifying its presentation. The tools used for creating and applyingabstraction patterns are shown along with an intelligence analysis example.

download:

• paper: (http://www.aaai.org/Papers/FLAIRS/2007/Flairs07-016.pdf)
• slides: