Automatic relation extraction

​IA161 Advanced NLP Course?, Course Guarantee: Aleš Horák

Prepared by: Adam Rambousek

State of the Art

References

Approx 3 current papers (preferably from best NLP conferences/journals, eg. ​ACL Anthology) that will be used as a source for the one-hour lecture:

1. Lefever, Els, Marjan Van de Kauter, and Véronique Hoste. "Evaluation of Automatic Hypernym Extraction from Technical Corpora in English and Dutch." Proceedings of the 9th International Conference on Language Resources and Evaluation (LREC 2014). 2014.
2. Wang, Tong, and Graeme Hirst. "Exploring patterns in dictionary definitions for synonym extraction." Natural Language Engineering 18.03 (2012): 313-342.
3. Schropp, Gwendolijn, Els Lefever, and Véronique Hoste. "A Combined Pattern-based and Distributional Approach for Automatic Hypernym Detection in Dutch." RANLP. 2013.
4. Grefenstette, Gregory. "INRIASAC: Simple Hypernym Extraction Methods." arXiv preprint arXiv:1502.01271 (2015).

Practical Session

Enhance hypernym detection to provide better results.

• Download prepared scripts and data
• Unzip, cd ia161-hyper and run ./hyper.py
• If you have trouble with encoding, use PYTHONIOENCODING=UTF-8 ./hyper.py
• The script reads file vstup.txt (each line is word|definition) and outputs hypernym for each word.
• Default approach is naive: first noun in definition is hypernym
• majka gives noun (k1) to some adjectives (k2), deal with this to improve results
• Update the find_hyper() function in hyper.py to provide better results.
• Upload updated script plus the output.
• Gold standard to evaluate your result: gold.txt