Exploiting query reformulations for web search result diversification

Abstract

When a Web user's underlying information need is not clearly specified from the initial query, an effective approach is to diversify the results retrieved for this query. In this paper, we introduce a novel probabilistic framework for Web search result diversification, which explicitly accounts for the various aspects associated to an underspecified query. In particular, we diversify a document ranking by estimating how well a given document satisfies each uncovered aspect and the extent to which different aspects are satisfied by the ranking as a whole. We thoroughly evaluate our framework in the context of the diversity task of the TREC 2009 Web track. Moreover, we exploit query reformulations provided by three major Web search engines (WSEs) as a means to uncover different query aspects. The results attest the effectiveness of our framework when compared to state-of-the-art diversification approaches in the literature. Additionally, by simulating an upper-bound query reformulation mechanism from official TREC data, we draw useful insights regarding the effectiveness of the query reformulations generated by the different WSEs in promoting diversity.

Select, Link and Rank: Diversified Query Expansion and Entity Ranking Using Wikipedia

“…Since then, there have been methods to explicitly model query aspects and diversify search results using query reformulations [20], query logs [11] and click logs [15], many of which use MMR-style diversification. Table 1.…”

Personalized diversification of search results

“…We take as starting point the original formulations of two representative intent-oriented diversification algorithms, namely IA-Select [1] and xQuAD [15,16], characterized by using an explicit representation of query intents for diversification. Other diversification schemes not explicitly using aquery aspect space [5,6,21,22] might be generalized as well towards personalized versions, which we envision as future work.…”

“…More recently, strands of research in the community considered other angles of the practical effectiveness of the output of retrieval systems, and the role that diversity, in particular, plays along with relevance in delivering effective value to the users of such systems [5,6,22]. The value of IR diversity is motivated by the uncertainty involved in a single query as the only evidence of the user information need [1,7,15,16]. The assumption is that different users may mean slightly (or even quite) different things by the same query expression.…”

“…Specifically, we introduce a user random variable in probabilistic diversification models previously developed in the literature on IR diversity [1,16]. We describe in detail a full approach to develop the resulting framework into computable components.…”

mNIR: Diversifying Search Results Based on a Mixture of Novelty, Intention and Relevance

“… xQuAD [16]: Santos et al [16] introduced the xQuAD probabilistic framework for search result diversification, which explicitly represented different query aspects as 'sub-queries'. They defined a diversification objective based on the estimated relevance of documents to multiple sub-queries, as well as on the relative importance of each sub-query to the original query.…”