Natural Language Processing
Semantic Interpretation and Knowledge Extraction
Since 2010 I have been working on a system that performs semantic interpretation of biomedical texts and extracts knowledge from those interpretations. There isn't a publication yet, but I prepared a presentation for an early version of the system in 2011. [PDF presentation]
In 2013, for a class project, I modified the system to incorporate some aspects of machine learning in order to learn some rules that I had previously been writing by hand. I'm no longer pursuing this line of research though I may in the future in a different way. The paper and presentation also include a more up-to-date summary of the system than the presentation above. [PDF paper] [PDF presentation]
Presentations and Summaries of Published Papers
I presented on/summarized the following papers during two courses on natural language understanding at UCF.
| Presentation | Original Paper |
|---|---|
| PDF, 2011 | Linking genes to literature: text mining, information extraction, and retrieval applications for biology. (Krallinger et al., 2008, Genome Biology) |
| PDF, 2011 | Extracting protein interactions from text with the unified AkaneRE event extraction system. (Saetre et al., 2010, IEEE/ACM Transaction on Computational Biology and Bioinformatics) |
| PDF, 2008 | Noun Homograph Disambiguation Using Local Context in Large Text Corpora. (Hearst, 1991, The Proceedings of the 7th Annual Conference of the UW Centre for the New OED and Text Research: Using Corpora, Oxford) |
| PDF, 2008 | Senselearner: Minimally supervised word sense disambiguation for all words in open text. (Mihalcea and Faruque, 2004, In Proceedings of ACL/SIGLEX Senseval-3) |
| Summary | Original Paper |
|---|---|
| PDF, 2008 | Automatic acquisition of hyponyms from large text corpora. (Hearst, 1992, COLING '92 Proceedings of the 14th conference on Computational linguistics - Volume 2) |
Knowledge-Lean Approaches to Word Sense Disambiguation (2009)
Download the paper or the presentation.
Abstract
This paper summarizes and discusses five different papers on knowledge-lean word sense disambiguation. Four of the papers have to do with Word Sense Induction (WSI), which is word sense disambiguation except with the additional task of discovering word senses automatically without relying on an external lexicon. The other paper is an example of word sense disambiguation that uses very little knowledge.
Presenting: A Brief Introduction to NLTK
In 2011, I prepared these slides to introduce AI students to using NLTK and Python for NLP.