Listwise approach to learning to rank: theory and algorithm

By Fen Xia, Tie-Yan Liu, Jue Wang, Wensheng Zhang, Hang Li. 2008

920 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Information Retrieval & Search

How researchers see this paper

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

In Artificial Intelligence

In Information Retrieval & Search

Citations in the corpus

Adaptive Information Seeking for Open-Domain Question Answering (2021, background): “Considering the belief state b = ⟨q, {p1, p2, · · · , p|C|}⟩, the ListMLE (Xia et al., 2008) ranking loss of the evidence scoring function φ is defined as the negative log likelihood of the ground truth permutation,”
VD-BERT: A Unified Vision and Dialog Transformer with BERT (2020, methodology): “We also show three other apparent attention patterns: attentions that a token puts to its previous token, to itself, and to the next token.”
A Neural Autoregressive Approach to Collaborative Filtering (2016, background): “Actually, from the perspective of learning-to-rank, CF-NADE acts as a ranking function which produces rankings of ratings based on previous ratings, where sjmoi (rmo<i ) corresponds to the score function in (Xia et al., 2008) and the rankings, ydown and yup, corresponds to true rankings that we woul”
LETOR: A benchmark collection for research on learning to rank for information retrieval (2010, neutral): “…et al. 2005; Freund et al. 2003; Herbrich et al. 1999; Matveeva et al. 2006; Qin et al. 2007; Tsai et al. 2007), and the listwise approach (Cao et al. 2007; Huang et al. 2009; Qin et al. 2008b, c, 2008d, e; Taylor et al. 2008; Volkovs et al. 2009; Xia et al. 2008; Xu et al. 2007; Yue et al. 2007).”
A Deep Look into Neural Ranking Models for Information Retrieval (2019, positive): “[98] proposed ListMLE for listwise ranking as”
Learning to Rank Using User Clicks and Visual Features for Image Retrieval (2015, background): “To deal with this problem, listwise approaches [11], [12] have been proposed to learn a ranking function by adopting separate lists as samples.”
The scoring of poses in protein-protein docking: current capabilities and future directions (2013, methodology): “Further, listwise ranking methods could be employed where, instead of viewing each pose independently and assigning it a score, and instead of comparing pairs of poses to determine which is superior, the whole list is ranked simultaneously as an inseparable set [137-142].”
EsdRank: Connecting Query and Documents through External Semi-Structured Data (2015, methodology): “We will ﬁrst revisit ListMLE [27], the listwise learning to rank model which Latent-ListMLE is built upon.”
On the Consistency of Ranking Algorithms (2010, methodology): “3 and Theorem 5 in Xia et al., 2008).”
A Short Introduction to Learning to Rank (2011, neutral): “The listwise approach includes ListNet [18], ListMLE [19], AdaRank [20], SVM MAP [21], and Soft Rank [22].”
Convex Calibration Dimension for Multiclass Loss Matrices (2014, background): “…in understanding consistency properties of surrogate risk minimization algorithms for various learning problems, such as binary clas-siﬁcation, multiclass 0-1 classiﬁcation, and various forms of ranking and multi-label prediction problems [18, 14, 34, 26, 2, 33, 28, 27, 8, 31, 10, 23, 4, 12, 15].”
A cross-benchmark comparison of 87 learning to rank methods (2015, neutral): “…2012; Sun, Huang, & Feng, 2011; Swersky, Tarlow, Adams, Zemel, & Frey, 2012; Wang & Xu, 2010; Wang, Kuai, Huang, Li, & Ni, 2008; Wu et al., 2011; Xia, Liu, Wang, Zhang, & Li, 2008; Xu, Chapelle, & Weinberger, 2012; Xu, Kersting, & Joachims, 2010; Zhu, Chen, et al., 2009; Zhou, Ding, You, &…”
Principled Reinforcement Learning with Human Feedback from Pairwise or K-wise Comparisons (2023, methodology): “Algorithms: MLE and Pessimistic MLE We first bound the estimation error for MLE, the most common algorithm in learning to rank and RLHF (Liu et al., 2009; Xia et al., 2008; Cao et al., 2007; Christiano et al., 2017a; Ouyang et al., 2022).”
Ranking Measures and Loss Functions in Learning to Rank (2009, methodology): “We have proved that the essential loss is both an upper bound of the measure-based ranking errors, and a lower bound of the loss functions in the aforementioned methods.”
Learning to Rank by Optimizing NDCG Measure (2009, methodology): “ListMLE [15] employs the likelihood loss as the surrogate for the IR evaluation metrics.”
On the Consistency of Multi-Label Learning (2011, positive): “…on optimizing a surrogate loss function has been well-studied for binary classiﬁcation (Zhang, 2004b; Steinwart, 2005; Bartlett et al., 2006), multi-class classiﬁcation (Zhang, 2004a; Tewari and Bartlett, 2007), learning to rank (Cossock and Zhang, 2008; Xia et al., 2008; Duchi et al., 2010), etc.”
Direct Learning to Rank and Rerank (2018, neutral): “There are also algorithms for ordinal regression , which is a related but different learning problem (Li et al, 2007; Crammer et al, 2001; Herbrich et al, 1999), and listwise approaches to ranking (Cao et al, 2007; Xia et al, 2008; Xu and Li, 2007; Yue et al, 2007).”