NIF workshop: An open benchmark for Wikipedia-Based NER
The NLP Interchange Format (NIF) 2.0 is currently created and is taking shape. One large topic is to create interoperable Named Entity Recognition and Linking tools and corpora to ease the engineering burden, when benchmarking such systems. This workshop is a developer meeting to talk about:
- NIF & NER: how currently NIF meets the needs of the NER community, and what is missing? Is it feasible to create a universal best practice?
- NIF & GATE: whether and how can NIF be incorporated in GATE?
- NIF & GATE NER eval framework: brainstorming on ideas for future research on how multilingual NER tools can be evaluated with support of GATE, DBpedia and possibly NIF.
Time and place
Wednesday the 9th of October. NB 473 (4th floor of the “New Building”), #2 on the campus plan. Room capacity is 38.
Participation and Registration
The workshop is free and open, so you can just show up to the workshop and join the discussion. We would kindly ask you to tell us and others that you will attend and add your name to this registration form.
- 9:00 – 9:45 – Sebastian, State-of-Play NIF
- 9:45 – 10:00 – Feedback time
- 10:00 – 10:30 – Milan and Tomas, GATE Evaluation Framework for Benchmarking Wikipedia-Based NER
- 10:30 – 10:45 – Feedback time
- 10:45 – 11:05 – Martin Popel, Treex NLP system
- 11:05 - 11:20 – Feedback time
- 11:20 - 11:40 – Pavel Smrž, Semantic Enrichment of Cultural Herritage Content: Experience from the DECIPHER Project
- 11:40 – 11:55 – Feedback time
- 11:55 – 12:15 – Discussion session (design of NER Benchmarking, guidelines)
- 12:15 – 14:00 – Lunch
- 14:00 – 16:00 – Break out sessions with different tasks / implementation (to be defined during the discussion session)
- 16:00 – 16:30 – Wrap-up of the workshop, future directions