Research
I started in research, teaching computers to find meaning in text. That work led to an open-source tool, around 25 papers and, eventually, a company.
My research
The PhD
My PhD at the University of Waikato, supervised by Professor Ian H. Witten, was in natural language processing and machine learning, and it was sponsored by Google. I focused on extracting meaning from text, and on how knowledge sources like Wikipedia can help computers understand what a document is about.
Work I'm known for
- Maui: an open-source tool for keyphrase extraction and topic indexing that extended the earlier KEA algorithm. It has been used by thousands of researchers, including a team at NATO, and cited in keyphrase extraction survey literature.
- Mining meaning from Wikipedia: a survey of how researchers use Wikipedia to extract meaning, published in the International Journal of Human-Computer Studies in 2009. It is my most-cited work, with around 410 citations.
- Human-competitive tagging using automatic keyphrase extraction: showed that automatic keyphrase extraction can match the quality of human taggers, EMNLP 2009.
In numbers
About 25 peer-reviewed publications, around 3,550 citations and an h-index of about 20. The full list of publications is also on Google Scholar and on dblp, where I appear as Olena Medelyan.
From research to industry
I interned on Google's sentiment analysis team in New York and was a finalist for the Google Australia and New Zealand Anita Borg Scholarship in 2008. After my PhD I was Chief Research Officer at Pingar, a text analytics company. I also ran a consultancy called Entopix before starting Thematic.