Source selection of long tail sources for federated search in an uncooperative setting

Urak, Günter Christoph; Ziak, Hermann; Kern, Roman

doi:10.1145/3167132.3167212

Cited by 4 publications

(6 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To examine the performance of the proposed model, we take all four categories of resource selection approaches as baselines, including lexicon-based (Taily [3]), sample-based (Rank-S [28], ReDDE [47], CRCS [45] and KBCS [19]), supervised (L2R [7] and Jnt [22]) and combination-based (SSLTS [49] and ECOMSVZ [24]). These baseline methods are briefly described below: KBCS [19] is one of the recently proposed methods in the samplebased category that has the best performance among methods of this category.…”

Section: Baselinesmentioning

confidence: 99%

“…It combines a number of strategies to rank resources for a given query, namely i) topical closeness between a query and resources, ii) estimation of a resource popularity from online sources, and iii) query expansion. SSLTS [49] is also a combination-based method and consists of all three strategies proposed in [24] and ranks resources by giving more weights to small (specialized) resources. Taily [3] is a method in the lexicon-based category.…”

Section: Baselinesmentioning

confidence: 99%

“…This metric measures the relevance scores of the top ranked k resources, normalized by the relevance scores of the best possible k resources for the given query. nP@{1,5} is used to compare with baselines SSLTS [49] and ECOMSVZ [24]. • Normalized Discounted Cumulated Gain at 𝑘 or nDCG@𝑘.…”

Section: Evaluation Metricsmentioning

confidence: 99%

“…They divided it into 100 resources, and evaluation queries were acquired from TREC Web Track 2009. Other two baseline methods SSLTS [49] and ECOMSVZ [24] used the FedWeb14 dataset, and for evaluation, judged queries were utilized.…”

Section: Datasetsmentioning

confidence: 99%

“…nDCG also discounts the value of a relevant resource when a resource is ranked lower. nDCG@{10,20} is used to compare with SSLTS [49] and ECOMSVZ [24].…”

Section: Evaluation Metricsmentioning

confidence: 99%

See 4 more Smart Citations

Learning To Rank Resources with GNN

Ergashev

Dragut

Meng

2023

Proceedings of the ACM Web Conference 2023

View full text Add to dashboard Cite

As the content on the Internet continues to grow, many new dynamically changing and heterogeneous sources of data constantly emerge. A conventional search engine cannot crawl and index at the same pace as the expansion of the Internet. Moreover, a large portion of the data on the Internet is not accessible to traditional search engines. Distributed Information Retrieval (DIR) is a viable solution to this as it integrates multiple shards (resources) and provides a unified access to them. Resource selection is a key component of DIR systems. There is a rich body of literature on resource selection approaches for DIR. A key limitation of the existing approaches is that they primarily use term-based statistical features and do not generally model resource-query and resource-resource relationships. In this paper, we propose a graph neural network (GNN) based approach to learning-to-rank that is capable of modeling resource-query and resource-resource relationships. Specifically, we utilize a pre-trained language model (PTLM) to obtain semantic information from queries and resources. Then, we explicitly build a heterogeneous graph to preserve structural information of query-resource relationships and employ GNN to extract structural information. In addition, the heterogeneous graph is enriched with resource-resource type of edges to further enhance the ranking accuracy. Extensive experiments on benchmark datasets show that our proposed approach is highly effective in resource selection. Our method outperforms the state-of-the-art by 6.4% to 42% on various performance metrics. CCS CONCEPTS• Information systems → Learning to rank; Combination, fusion and federated search.

show abstract

Section: Baselinesmentioning

confidence: 99%

Section: Baselinesmentioning

confidence: 99%

Section: Evaluation Metricsmentioning

confidence: 99%

Section: Datasetsmentioning

confidence: 99%

“…nDCG also discounts the value of a relevant resource when a resource is ranked lower. nDCG@{10,20} is used to compare with SSLTS [49] and ECOMSVZ [24].…”

Section: Evaluation Metricsmentioning

confidence: 99%

See 3 more Smart Citations