Lubell-Doughtie and Hofmann on Learning to Rank from Relevance Feedback for e-Discovery

Peter Lubell-Doughtie of Columbia University and Katja Hofmann of the Intelligent Systems Lab Amsterdam (ISLA) at the University of Amsterdam, presented a paper entitled Learning to Rank from Relevance Feedback for e-Discovery, at ECIR 2012: The European Conference on Information Retrieval, held 1-5 April 2012 in Barcelona, Catalonia, Spain. Here is the abstract:

In recall-oriented search tasks retrieval systems are privy to a greater amount of user feedback. In this paper we present a novel method of combining relevance feedback with learning to rank. Our experiments use data from the 2010 TREC Legal Track to demonstrate that learning to rank can tune relevance feedback to improve result rankings for specific queries, even with limited amounts of user feedback.

Cormack et al., Overview of the TREC 2010 Legal Track

Professor Dr. Gordon V. Cormack of the University of Waterloo; Maura R. Grossman, Esq., of Wachtell, Lipton; Bruce Hedin of H5, and Professor Dr. Douglas W. Oard of the University of Maryland, have posted Overview of the TREC 2010 Legal Track. Here is the abstract:

TREC 2010 was the fifth year of the Legal Track, which focuses on evaluation of search technology for discovery of electronically stored information in litigation and regulatory settings. The TREC 2010 Legal Track consisted of two distinct tasks: the Learning task, in which participants were required to estimate the probability of relevance for each document in a large collection, given a seed set of documents, each coded as responsive or non-responsive; and the Interactive task, in which participants were required to identify all relevant documents using a human-in-the-loop process.