Predicting query performance for fusion-based retrieval

Markovits, Gad; Shtok, Anna; Kurland, Oren; Carmel, David

doi:10.1145/2396761.2396865

Cited by 18 publications

(15 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Markovits et al employ a simpler variant of WIG in a data fusion setting [8]. In this variant, called here ScoreAvg, instead of using rel(C, qi) for normalization as in WIG, the relevance scores rel(d, qi) are sum normalized to [0, 1] before computing their average.…”

Section: Baseline Qpps For Aspect Weightingmentioning

confidence: 99%

“…ScoreDev. This method [8] is a variant of NQC, and applies Eq. 2 without the normalization factor rel(C, qi).…”

Section: Baseline Qpps For Aspect Weightingmentioning

confidence: 99%

“…To this end, we employ post-retrieval QPPs that are based on score distribution analysis, namely, weighted information gain (WIG) [14], normalized query commitment (NQC) [12] and their variants presented in [8]. The choice of these QPPs is intentional, to satisfy the demanding efficiency requirements of online query processing.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Query Performance Prediction for Aspect Weighting in Search Result Diversification

Ozdemiray

Altıngövde

2014

Proceedings of the 23rd ACM International Conference on Conference on Information and Knowledge Management

View full text Add to dashboard Cite

Accurate estimation of query aspect weights is an important issue to improve the performance of explicit search result diversification algorithms. For the first time in the literature, we propose using post-retrieval query performance predictors (QPPs) to estimate, for each aspect, the retrieval effectiveness on the candidate document set, and leverage these estimations to set the aspect weights. In addition to utilizing well-known QPPs from the literature, we also introduce three new QPPs that are based on score distributions and hence, can be employed for online query processing in reallife search engines. Our exhaustive experiments reveal that using QPPs for aspect weighting improves almost all stateof-the-art diversification algorithms in comparison to using a uniform weight estimator. Furthermore, the proposed QPPs are comparable or superior to the existing predictors in the context of aspect weighting.

show abstract

Section: Baseline Qpps For Aspect Weightingmentioning

confidence: 99%

“…ScoreDev. This method [8] is a variant of NQC, and applies Eq. 2 without the normalization factor rel(C, qi).…”

Section: Baseline Qpps For Aspect Weightingmentioning

confidence: 99%

See 1 more Smart Citation

Query Performance Prediction for Aspect Weighting in Search Result Diversification

Ozdemiray

Altıngövde

2014

Proceedings of the 23rd ACM International Conference on Conference on Information and Knowledge Management

View full text Add to dashboard Cite

show abstract

“…To estimate p(d|Iq) using information induced from the intermediate lists, and inspired by some recent work on predicting query-performance for fusion [25], we can write…”

Section: Exploiting the Special Characteristics Of The Fusion Settingmentioning

confidence: 99%

Utilizing relevance feedback in fusion-based retrieval

Rabinovich

Rom²,

Kurland

2014

Proceedings of the 37th International ACM SIGIR Conference on Research &Amp; Development in Information Retrieval

Self Cite

View full text Add to dashboard Cite

Work on using relevance feedback for retrieval has focused on the single retrieved list setting. That is, an initial document list is retrieved in response to the query and feedback for the most highly ranked documents is used to perform a second search. We address a setting wherein the list for which feedback is provided results from fusing several intermediate retrieved lists. Accordingly, we devise methods that utilize the feedback while exploiting the special characteristics of the fusion setting. Specifically, the feedback serves two different, yet complementary, purposes. The first is to directly rank the pool of documents in the intermediate lists. The second is to estimate the effectiveness of the intermediate lists for improved re-fusion. In addition, we present a meta fusion method that uses the feedback for these two purposes simultaneously. Empirical evaluation demonstrates the merits of our approach. As a case in point, the retrieval performance is substantially better than that of using the relevance feedback as in the single list setting. The performance also substantially transcends that of a previously proposed approach to utilizing relevance feedback in fusion-based retrieval.

show abstract

“…In this paper, we follow Arguello et al [3] and also use one copy of each source feature for every source, which yields better performance according to our experiments as well. Then, to solve our problem of blending results, we propose a pair-wise approach inspired by [3,15], but adapted to our problem as well as a point-wise approach similar in aim to the method described in [19], where it was outperformed by Round-Robin [21], and a list-wise approach that is the most fitted to our problem and allows us to use features from [12], where query performance was predicted.…”

Section: Related Workmentioning

confidence: 99%

Blending Vertical and Web Results

Lefortier

Serdyukov

Romanenko

et al. 2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Modern search engines aggregate results from specialized verticals into the Web search results. We study a setting where vertical and Web results are blended into a single result list, a setting that has not been studied before. We focus on video intent and present a detailed observational study of Yandex's two video content sources (i.e., the specialized vertical and a subset of the general web index) thus providing insights into their complementary character. By investigating how to blend results from these sources, we contrast traditional federated search and fusion-based approaches with newly proposed approaches that significantly outperform the baseline methods.

show abstract

Predicting query performance for fusion-based retrieval

Cited by 18 publications

References 34 publications

Query Performance Prediction for Aspect Weighting in Search Result Diversification

Query Performance Prediction for Aspect Weighting in Search Result Diversification

Utilizing relevance feedback in fusion-based retrieval

Blending Vertical and Web Results

Contact Info

Product

Resources

About