Translating Queries into Snippets for Improved Query Expansion

Proceedings of the 22nd International Conference on Computational Linguistics (Coling 2008), pages 737–744 Manchester, August 2008 Translating Queries into Snippets for Improved Query Expansion Stefan Riezler and Yi Liu and Alexander Vasserman Google Inc. 1600 Amphitheatre Parkway Mountain View, CA 94043 {riezler,yliu,avasserm}@google.com Abstract User logs of search engines have recently been applied successfully to improve var- ious aspects of web search quality. In this paper, we will apply pairs of user queries and snippets of clicked results to train a machine translation model to bridge the “lexical gap” between query and document space. We show that the combination of a query-to-snippet translation model with a large n-gram language model trained on queries achieves improved contextual query expansion compared to a system based on term correlations. 1 Introduction In recent years, user logs of search engines have at- tracted considerable attention in research on query clustering, query suggestions, query expansion, or general web search. Besides the sheer size of these data sets, the main attraction of user logs lies in the possibility to capitalize on users’ input, either in form of user-generated query reformulations, or in form of user clicks on presented search results. However noisy, sparse, incomplete, and volatile these data may be, recent research has presented impressive results that are based on simply taking the majority vote of user clicks as a signal for the relevance of results. In this paper we will apply user logs to the prob- lem of the “word mismatch” or “lexical chasm” (Berger et al., 2000) between user queries and documents. The standard solution to this prob- lem, query expansion, attempts to overcome this c⃝ 2008. Licensed under the Creative Commons Attribution-Noncommercial-Share Alike 3.0 Unported li- cense (http://creativecommons.org/licenses/by-nc-sa/3.0/). Some rights reserved. mismatch in query and document vocabularies by adding terms with similar statistical properties to those in the original query. This will increase the chances of matching words in relevant documents and also decrease the ambiguity of the overall query that is inherent to natural language. A suc- cessful approach to this problem is local feed- back, or pseudo-relevance feedback (Xu and Croft, 1996), where expansion terms are extracted from the top-most documents that were retrieved in an initial retrieval round. Because of irrelevant results in the initial retrieval, caused by ambiguous terms or retrieval errors, this technique may cause expan- sion by unrelated terms, leading to query drift. Fur- thermore, the requirement of two retrieval steps is computationally expensive. Several approaches have been presented that de- ploy user query logs to remedy these problems. One set of approaches focuses on user reformu- lations of queries that differ only in one segment (Jones et al., 2006; Fonseca et al., 2005; Huang et al., 2003). Such segments are then identified as candidate expansion terms, and filtered by var- ious signals such as cooccurrence in similar ses- sions or log-likelihood ratio of original and ex- pansion phrases. Other approaches focus on the relation of queries and retrieval results, either by deploying the graph induced by queries …

Content truncated to 500 words