Beyond independent relevance: methods and evaluation metrics for subtopic retrieval

By Chengxiang Zhai, William W. Cohen, John D. Lafferty. 2003

571 citations recorded by the source corpus.

Subfields: Information Systems, Management Science and Operations Research

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Information Systems

In Management Science and Operations Research

Citations in the corpus

Determinantal Point Processes for Machine Learning (2012, neutral): “…a number of recently developed techniques for working with diverse sets, particularly in the information retrieval community [Carbonell and Goldstein, 1998, Zhai et al., 2003, Chen and Karger, 2006, Yue and Joachims, 2008, Radlinski et al., 2008, Swaminathan et al., 2009, Raman et al., 2012].”
Relevance-based Word Embedding (2017, background): “However, this objective is not necessarily equivalent to the main objective of many IR tasks. primary objective in many IR methods is to model the notion of relevance [20, 34, 43].”
Ranking with Fairness Constraints (2017, background): “…w ij s to attempt to capture the amount of diversity item i would introduce at position k conditioned on the items that were placed at positions 1 , . . . , k − 1 (see [CG98, ZH08, CKC + 08, ZCL03, ZMKL05]), or casting it directly as an (unconstrained) multi-objective optimization problem [YS16].”
Test Collection Based Evaluation of Information Retrieval Systems (2010, background): “[293] defined a diversity-specific version of precision and formalized Over’s definition of aspectual recall, choosing to instead to call it sub-topic recall (S-recall).”
Diversifying search results (2009, background): “Zhai et al [19] point out that it is in general insufficient to simply return a set of relevant results.”
Online learning to diversify from implicit feedback (2012, background): “[2, 19, 3, 16, 4]), recently developed supervised learning methods for diversity have shown a lot of promise (e.”
Novelty and diversity in information retrieval evaluation (2008, methodology): “Zhai and colleagues [31, 32] develop and validate subtopic retrieval methods based on a risk minimization framework and introduce corresponding measures for subtopic recall and precision.”
Less is More Probabilistic Models for Retrieving Fewer Relevant Documents (2006, methodology): “Table 2 lists the instance recall at rank ten results, along with instance recall scores computed on result sets from the subtopic [21].”
Document Comprehensiveness and User Preferences in Novelty Search Tasks (2015, background): “The simplest metric that accounts for different subtopics is subtopic recall [12].”
Assessor Differences and User Preferences in Tweet Timeline Generation (2015, methodology): “The evaluation metrics for TTG represent straightforward extensions of previous work: aspect recall [21], sub-topic re-call [34], and the“nugget pyramid”approach from the TREC question answering evaluations [17].”
Overview of the TREC-2014 Microblog Track (2014, methodology): “The tweet timeline generation task represents a natural extension of classic ad hoc retrieval and shares similarities with previous tasks such as aspect retrieval [10], sub-topic retrieval [18], and the notion of “information nuggets” in TREC question answering evaluations [15, 3].”
Exploiting query reformulations for web search result diversification (2010, background): “, cosine), Zhai and Lafferty [33] proposed to model relevance and redundancy within the language modelling framework.”
An axiomatic approach for result diversification (2009, methodology): “We note that this measure is same as the Srecall measure proposed in [21].”
Beliefs and biases in web search (2013, background): “Research on result diversity (e.g., [25][40][48]) is also relevant but typically focuses on aspects such as topical variance, rather than results skewed toward one perspective or biased search behaviors.”
Learning to rank for information retrieval (2009, unknown): “Some of them [118] take the proximity of the query terms into consideration, and some others consider the relationship between documents in terms of content similarity [117], hyperlink structure [113], website structure [101], and topic diversity [139].”
Learning diverse rankings with multi-armed bandits (2008, methodology): “All previous approaches of which we are aware that optimize diversity similarly require a relevance function to be learned prior to performing a diversification step (Zhu et al., 2007; Zhang et al., 2005; Zhai et al., 2003), with the exception of Chen and Karger (2006).”