Simon Overell's
Publications

Posters

Distribution of Location References in Wikipedia
The Future of Multimedia Knowledge Management 2008, Milton Keynes

This poster will present our work mining location references from different language versions of Wikipedia. The extracted events will be visualised on a map. We will demonstrate that despite Wikipedia’s best efforts for neutrality, cultural biases are introduced. This analysis of Wikipedia will be of interest to the growing community of researches using Wikipedia as a pool of "World Knowledge." By analysing different language versions of Wikipedia we can show how different events have significance to different cultures. Finally we calculate a “bias index” for each language version of Wikipedia we analyse, this is the ratio of references to locations within countries where this is the native language spoken to references outside these countries.

@inproceedings{overell08a,
title={ Distribution of Location References in Wikipedia },
author={Simon Overell and Stefan R\"uger},
year={2008},
month={February},
booktitle = {MMKM Workshop: The Future of Multimedia Knowledge Management},
location = {Milton Keynes, UK},
pages = {23}
} 
SIRIL: A multidimensional browsing framework
MMKM Workshop 2007, Milton Keynes

SIRIL is part of the multidimensional browsing framework currently being developed by the MMIS team. It consists of an XML API, a dynamically generated user interface and a search engine supporting text, image & geographic IR. The framework will support multi-document types and browsing methods; any server should support any front-end. SIRIL will show case the MMIS term’s work on image indexing & browsing, image annotation and GIR.

@inproceedings{overell07a,
title={ {SIRIL}: {A} multidimensional browsing framework },
author={Peter Howarth and Jo\~ao Magalh\~aes and Simon Overell and Stefan R\"uger and Alexei Yavlinsky},
year={2007},
month={January},
booktitle = {MMKM Workshop: Multimedia Knowledge Managment: Industry meets academia},
location = {Milton Keynes, UK},
pages = {17}
} 

Simon Overell's Publications
About.me | Academia | Linked in | Publications | Stuff I've Built | Musings | Follow Me
My PhD topic was Geographic Information Retrieval. I've written papers on Geographic Disambiguation and Modelling, Patents on Classification and Accurate NLP at Scale and given talks on Extracting Data from Wikipedia and the Web. For abstracts and citation details on all my publications click the boxes below.

Theses

PhD Thesis. Geographic Information Retrieval: Classification, Disambiguation and Modelling. (Imperial College London, 2009)

Master’s Thesis. TRIDE: Implementation of a Teleo-Reactive Integrated Development Environment. (Imperial College London, 2005)

Journal Articles

View of the world according to Wikipedia: Are we all little Steinbergs? (JOCS, 2011)

Using co-occurrence models for placename disambiguation. (IJGIS, 2008)

Conference & Workshop Papers

Classifying Tags using Open Content Resources. (WSDM, 2009, Barcelona)

Geographic Co-occurrence as a Tool for GIR. (GIR @ CIKM, 2007, Lisbon)
...

Invited Talks

I've given 9 invited talks covering my PhD, research at Yahoo! and work at True Knowledge.

Invited Articles

The Problem of Place Name Ambiguity (The SIGSPATIAL Special, 2011)

Are we getting it right? The results of the Student Survey (Informer, Spring 2008)

Patents

I've written a various patents all broadly related to classification. Four have been granted with previous employers and two are pending with Spider.io.

Evaluation Conference Papers

A key part of Information Retrieval is evaluation. Due to the efforts of the TREC and CLEF conferences there are now a series of standardised data sets for these evaluations. I've taken part in three CLEF conferences and one TREC conference, publishing 10 papers.

Posters

Distribution of Location References in Wikipedia (The Future of Multimedia Knowledge Management 2008, Milton Keynes)

SIRIL: A multidimensional browsing framework (MMKM Workshop 2007, Milton Keynes)

Citations

Both Google Scholar and Microsoft Academic Search maintain co-author and citation lists.