LETOR: Benchmark Dataset for Research on Learning to Rank for Information Retrieval

By Tie-Yan Liu, Jun Xu, Tao Qin, Wen-Ying Xiong, Hang Li. 2007

527 citations recorded by the source corpus.

How researchers see this paper

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

1. Yoshua Bengio
2. Quoc V. Le
3. Dacheng Tao
4. Yi Yang
5. Sergey Levine
6. Mohit Bansal
7. Zhe Gan
8. Trevor Darrell

Citations in the corpus

• Learning to Branch in Mixed Integer Programming (2016, background): “This type of normalization is referred to as query-based normalization in the IR literature (Liu et al. 2007).”
• Test Collection Based Evaluation of Information Retrieval Systems (2010, methodology): “The first shared LTR data set was the LETOR benchmark [168].”
• Learning to Rank by Optimizing NDCG Measure (2009, methodology): “To study the performance of NDCG Boost we use the latest version (version 3.0) of LETOR package provided by Microsoft Research Asia [22].”
• Tighter Bounds for Structured Estimation (2008, methodology): “We compared the results of our method and two standard methods for ranking: ranking SVM [10, 8] and RankBoost [6] (the baselines for OHSUMED are shown in [13]) and used NDCG as the performance criterion.”
• A Short Introduction to Learning to Rank (2011, unknown): “Benchmark data sets on learning to rank have also been released [4].”
• An Introduction to Neural Information Retrieval (2018, unknown): “, [Liu et al., 2007]) that don’t contain the textual contents.”
• Learning to rank for information retrieval (2009, unknown): “However, there were no such benchmark datasets for ranking until the LETOR collection [79] was released in early 2007.”
• Click Models for Web Search (2015, background): “is includes, but is not limited to, the TREC datasets [Voorhees and Harman, 2001], the Microsoft LETOR datasets [Liu et al., 2007], the Yahoo! Learning to Rank Challenge3 and the Yandex Internet Mathematics Challenge.”
• Listwise approach to learning to rank: theory and algorithm (2008, background): “One data set is synthetic data, and the other is the LETOR benchmark data for learning to rank (Liu et al., 2007).”
• Learning to rank for quantity consensus queries (2009, background): “We defined two families of features on (the query and) snippet text: first, standard vector-space ranking features [14, 13], and second, features that encode lexical proximity between query word matches and quantity tokens [16, 6, 2].”
• Efficient algorithms for ranking with SVMs (2010, methodology): “We experiment with the Ohsumed and TD 2004 datasets from the Letor 3.0 distri- bution 4 (Liu et al, 2007).”
• A general approximation framework for direct optimization of information retrieval measures (2010, methodology): “We used LETOR datasets (Liu et al. 2007) in our experiments.”
• BoltzRank: learning to maximize expected ranking gain (2009, methodology): “In this section we describe the experiments that we carried out on LETOR3.0 OHSUMED and TD2004 ( Liu et al. 2007 ) data sets.”
• Learning to Rank for Information Retrieval and Natural Language Processing (2011, unknown): “, [70]), and a competition on the task 3 has also been carried out.”
• Ranking Measures and Loss Functions in Learning to Rank (2009, methodology): “From the definitions of NDCG and MAP, we can see that their maximum values are both one.”