Filtering Personal Queries from Mixed-Use Query Logs

@incollection{ BressaneNeto_al:2014,
booktitle={Advances in Artificial Intelligence},
series={Lecture Notes in Computer Science},
editor={Sokolova, Marina and Beek, Peter},
title={Filtering Personal Queries from Mixed-Use Query Logs},
publisher={Springer International Publishing},
author={Bressane Neto, Ary Fagundes and Desaulniers, Philippe and Duboue, Pablo Ariel and Smirnov, Alexis},

This paper documented research done at a customer site. It received a best paper award at the Canadian AI 2014, Montreal.

It addresses the problem of text mining employee search queries while respecting their expectations of privacy, as they can use their work computers for personal matters during breaks. Similar to ComplementNB, we found it was much performant to consider their work queries as the signal and their personal queries (much varied in nature) as the background.

On a personal note, as I contributed most of the writing, the fact that it got a best paper award distinction encouraged me to pursue larger writing endeavors, culminating with my feature engineering book.

The paper is currently paywalled, but contact me if you would like a preprint.