120 likes | 127 Views
Get an overview of the IPCWLM project, IPCRECLASS, IPCPUB, and IPCCAT-neural cross lingual prototype discussed during the 39th session of the IPC Revision Working Group.
E N D
Report on IPC-related IT systems 39th session of the IPC Revision Working Group Geneva April 26, 2018 Patrick Fiévet Head, IT Systems Section International Classifications and Standards Division
Agenda • IPCWLM project / IPCRECLASS • IPCPUB • IPCCAT-neural cross lingual prototype
IPCWLM Project Status • IPCWLMS implementation contract and administrative process: • Contract signature: delayed (new target May 1, 2018) • Information security requirements clarification in progress • Planning to be revised at the end of the clarification period
IPCWLM Project Status • Requirements clarification: started • Outcome of the IPCWLM Task Force shared with contractor • WL simulation report: specification ready to be proposed for approval • Analysis of infrastructure and technology aspects • Analysis of DOCDB XML fields required by IPCWLMS
IPCRECLASS • Stop of email-based submission of RL as from July 1, 2018 • New feature (imminent move into production): • Possibility to filter by Docdb patent family identifier • Interactive reclassification: • Occasional issue when adding a new symbol: Scrollbar added
IPCPUB- Status for version 7.6 • What is new since IPC CE50: • Order of search results is now according to selected search features e.g. if STATS search is used, STATS results appear first. • IPCCAT search: possibility to Expand the query box
IPCCAT-neural cross lingual prototype • Test with 1000 randomly selected patents in AR, DE, ES, FR, JA, RU, ZH • Difficult to compare, not the same distribution of patents
IPCCAT-neural cross lingual prototype • Test with 500 randomly selected patents from subclass G06F • Losses due to translation are more visible for each language • Needs to be evaluated for all languages
IPCCAT-neural cross lingual prototype-potential use in IPC reclassification • Next steps: • Simulate what cross-lingual text categorization would have done on a past IPC revision instead of Default Transfer • Proposed an approach and estimate resources for implementation • Report and conclusions for CE 51 decision
Incentive to R&D in text categorization: WIPO-Delta training collection • Incentives for research and development institutes interested in automatic text categorization : • WIPO DELTA 2018 EN collection available upon request • Fully specified XML format • 50 million documents classified in the IPC • Complement the public WIPO-ALPHA training collection • http://www.wipo.int/classifications/ipc/en/ITsupport/Categorization/dataset/index.html
IT operations and support for IPC Questions? Thank you