PARC Natural Language Processing: Examples

Natural Language Processing: Examples

Abstract Knowledge Representation

PARC's prototype system correctly answers a question based on a text passage. Despite the differences in the passage and question text strings, the system intelligently determines that the answer is YES – because it understands the relation between "acquired"/"buy", "International Business Machines"/"IBM", and "seven"/"any" as used in these text strings.

Detailed Abstract Knowledge Representations (AKR) are shown below. From these, the system can automatically generate similar texts with the same meaning and align them using entailment and contradiction logic to yield the correct answer above.
Detailed AKR of the passage:		Detailed AKR of the question:

view larger

BUSINESS CONTACT

Lawrence Lee
Director of Business Development, Intelligent Systems Laboratory
650-812-4756

SELECTED PAPERS

Exploiting F-structure input for sentence condensation, Proceedings of LFG Conference

Speed and accuracy in shallow and deep stochastic parsing, Human Language Technology Conference/ Association for Computational Linguistics

Packed rewriting for mapping semantics to KR, International Workshop on Computational Semantics

Parsing the Wall Street Journal using a lexical-functional grammar and discriminative estimation techniques, Proceedings of the Annual Meeting of the Association for Computational Linguistics

Other Resources:

Lexical Functional Grammar

info@parc.com
privacy policy