U Albany & USC at the TREC 2012 Session Track
Xiaojun Yuan, Jingling Liu, Ning Sa · Text REtrieval Conference · 2012
Research in Information Retrieval has noticed that it is often the case that in a search process, users begin an interaction with a sufficiently under-specified query and they will need to reformulate their query multiple times before they find desired information. In terms of this, researchers hypothesized that a search engine may be able to better serve a user “by ranking results that help “point the way” to what the user is really looking for, or by complementing results from previous queries in the sequence with new results, or in other currentlyunanticipated ways.” (Kanoulas, Carterettey, Hallz, Cloughx, & Sanderson, 2011). In 2012, the goal of session track is: (G1) to test whether system performance can be improved for a given query by using previous queries and user interactions with the system (including clicks on ranked results, dwell times, etc.), and (G2) to evaluate system performance over an entire query session instead of a single query. Our UAlbany and USC group joined the Session Track task by taking into account searcher behaviors during the course of their information seeking process. In the following, we give an overview of our approach, details of our submission runs, our results, and our conclusions about the results.