Natural Language Analysis of Patent Claims

1
Natural Language Analysis of Patent Claims

• Svetlana Sheremetyeva
• Department of Computational Linguistics
• Copenhagen Business School
• Denmark

2
Overview of the presentation

• Why do we need NLP of patent claims
• Natural language analyzer for patent claims
• Examples of applications
• Conclusions

3
Why NLP of patent claims

• The claim is the focal point of a patent
  disclosure and the actual subject of legal
  protection
• An optimistic view is that NLP of patent claims
  will raise a level of performance of patent
  related applications
• Patent claims are particularly challenging for
  NLP as they are an ultimate example of extremely
  long sentences frequently including a lot of
  telescopically embedding clauses.

4
Example of a US Patent Claim
No existing NLP system, can process patent texts
(claims) adequately.

• A cassette for holding excess lengths of light
  waveguides in a splice area comprising
• a cover part and a pot-shaped bottom part
  having a bottom disk and a rim extending
  perpendicular to said bottom disk, said cover and
  bottom parts are superimposed to enclose
  jointly an area forming a magazine for excess
  lengths of light waveguides, said cover part
  being rotatable in said bottom part, two guide
  slots formed in said cover part, said slots being
  approximately radially directed, guide members
  disposed on said cover part, a splice holder
  mounted on said cover part to form a rotatable
  splice holder.

5
Natural language analyzer for patent claims
(knowledge)

• Lexicons
• shallow lexicon for supertagging
• deep predicate lexicon
• Grammar mixture of PHG and DG
• Knowledge representation (final parse) in two
  formats
• a set of individual predicate-argument structures
• a tree of predicate argument structures

6
Natural language analyzer for patent claims
(analysis algorithm)

• Tokenazation
• Supertaging (shallow lexicon)
• assignment of supertags, disambiguation
• Chunking (keeps internal structure of chunks)
• Determining dependencies (predicate lexicon)
• case-role dependencies predicate disambiguation
• predicate-argument structures

7
A claim text chunked into phrases
8
Final parse as a set of predicate-argument
structures
9
Examples of applications

• Machine translation
• Transfer of SL predicate-argument structures into
  TL predicate argument structures
• Generation of TL claim with the generator of the
  application for authoring patent claims
  (Sheremetyeva, 2003)
• AutoRead for improving the readability of claims
• Generation of the simple sentences based on every
  predicate structure

10
A screenshot of the AutoRead prototype interface
11
Conclusions

• Suggested methodology can be used in any patent
  related application including MT, IR, QA, etc. in
  multiple languages
• We intend to
• add an optional interactive module to the
  analyzer to improve the quality of analysis
• integrate the analyzer into a number of
  applications, e.g., MT, AutoRead, IR, QA, etc.
• develop such applications in multiple languages