Entity Linking
Download as PPTX, PDF
0 likes221 views
The document discusses entity linking, which involves connecting natural language references to real-world entities. It first reviews existing entity recognition and relationship modeling techniques. It then examines earlier "brute force" entity linking methods based on string and context matching. The document outlines some improvements to these methods, such as hashing techniques. Finally, it discusses more advanced approaches that use collective contexts and knowledge graph statistics to better handle challenging cases.
Entity Linking
- 3. Entity Relationship Model https://en.wikipedia.org/wiki/Entity%E2%80%93relationship_model#/media/File:ER_D iagram_MMORPG.png Chen et al., 1997
- 4. NER (Named Entity Recognition) http://brat.nlplab.org/img/examples/train.txt-doc-109-small.png (Example of NP chunker output)
- 6. Brute Force Tactics faster system leverage linking features (Lin et al.) String/context matching Rao et al. Exact matches Partial matches Abbreviations Aliases Similarity
- 8. Brute Force Tactics contd. - Improvements Hashing vs. Linear Passes Similarity Calculations Dice Scoring
- 9. When force doesnt work http://www.webdesigndev.com/wp-content/uploads/2015/03/121.jpg
- 10. Better Tactics?
- 11. Vinculum (Ling et al.)
- 12. More Improvements Candidate Listing Collective Contexts (Lin et al.) WikiGraph Statistics ranking (Rao et al.)
- 13. Edge Case(s) Unlinkable Entities Use Google Books N-grams corpus for differentiation
- 14. Works Cited Peter Pin-Shan Chen. English, Chinese and ER Diagrams, Data and Knowledge Eng. 23 (1997) 5-16. Thomas Lin, Mausam, Oren Etzioni. Entity Linking at Web Scale. Delip Rao, Paul McNamee, Mark Dredze. Entity Linking: Finding Extracted Entities in a Knowledge Base. Xiao Ling, Sameer Singh, Daniel S. Weld. Design Challenges for Entity Linking.