Home

A project for developing tools to measure various types of word similarity and measure them on various datasets.

Subprojects / Division of labour (tentative)

0/a Acquire existing datasets and embeddings

0/b Build OpenRoget dataset

1. Build small ML framework for measuring word similarity / synonym detection
2. Measure all similarities on all datasets
3. Develop 4lang similarity

Technical

• Progress should be documented on this wiki
• Discussions should take place on the mailing list [email protected]
• Datasets should be stored under nessi6:/mnt/store/home/hlt/wordsim

Meetings, milestones

• Intro to 4lang similarity: February 3rd, 10.30 am, SZTAKI Rm 506
• ACL short paper deadline: Feb 29th, long paper: March 18th, StarSem: April 18th

Bibliography

• Hill et al. 2015: SimLex-999: Evaluating Semantic Models With (Genuine) Similarity Estimation