Tie-Yan Liu

Learning to Rank for Information Retrieval

Learning to rank for Information Retrieval (IR) is a task to automatically construct a ranking model using training data, such that the model can sort new objects according to their degrees of relevance, preference, or importance. Many IR problems are by nature ranking problems, and many IR technologies can be potentially enhanced by using learning-to-rank techniques. The objective of this tutorial is to give an introduction to this research direction. Specifically, the existing learning-to-rank algorithms are reviewed and categorized into three approaches: the pointwise, pairwise, and listwise approaches. The advantages and disadvantages with each approach are analyzed, and the relationships between the loss functions used in these approaches and IR evaluation measures are discussed. Then the empirical evaluations on typical learning-to-rank methods are shown, with the LETOR collection as a benchmark dataset, which seems to suggest that the listwise approach be the most effective one among all the approaches. After that, a statistical ranking theory is introduced, which can describe different learning-to-rank algorithms, and be used to analyze their query-level generalization abilities. At the end of the tutorial, we provide a summary and discuss potential future work on learning to rank.

Listwise approach to learning to rank: theory and algorithm

This paper aims to conduct a study on the listwise approach to learning to rank. The listwise approach learns a ranking function by taking individual lists as instances and minimizing a loss function defined on the predicted list and the ground-truth list. Existing work on the approach mainly focused on the development of new algorithms; methods such as RankCosine and ListNet have been proposed and good performances by them have been observed. Unfortunately, the underlying theory was not sufficiently studied so far. To amend the problem, this paper proposes conducting theoretical analysis of learning to rank algorithms through investigations on the properties of the loss functions, including consistency, soundness, continuity, differentiability, convexity, and efficiency. A sufficient condition on consistency for ranking is given, which seems to be the first such result obtained in related research. The paper then conducts analysis on three loss functions: likelihood loss, cosine loss, and cross entropy loss. The latter two were used in RankCosine and ListNet. The use of the likelihood loss leads to the development of a new listwise method called ListMLE, whose loss function offers better properties, and also leads to better experimental results.

LETOR: A benchmark collection for research on learning to rank for information retrieval

LETOR is a benchmark collection for the research on learning to rank for information retrieval, released by Microsoft Research Asia. In this paper, we describe the details of the LETOR collection and show how it can be used in different kinds of researches. Specifically, we describe how the document corpora and query sets in LETOR are selected, how the documents are sampled, how the learning features and meta information are extracted, and how the datasets are partitioned for comprehensive evaluation. We then compare several state-of-the-art learning to rank algorithms on LETOR, report their ranking performances, and make discussions on the results. After that, we discuss possible new research topics that can be supported by LETOR, in addition to algorithm comparison. We hope that this paper can help people to gain deeper understanding of LETOR, and enable more interesting research projects on learning to rank and related topics.

Support vector machines classification with a very large-scale taxonomy

Very large-scale classification taxonomies typically have hundreds of thousands of categories, deep hierarchies, and skewed category distribution over documents. However, it is still an open question whether the state-of-the-art technologies in automated text categorization can scale to (and perform well on) such large taxonomies. In this paper, we report the first evaluation of Support Vector Machines (SVMs) in web-page classification over the full taxonomy of the Yahoo! categories. Our accomplishments include: 1) a data analysis on the Yahoo! taxonomy; 2) the development of a scalable system for large-scale text categorization; 3) theoretical analysis and experimental evaluation of SVMs in hierarchical and non-hierarchical settings for classification; 4) an investigation of threshold tuning algorithms with respect to time complexity and their effect on the classification accuracy of SVMs. We found that, in terms of scalability, the hierarchical use of SVMs is efficient enough for very large-scale classification; however, in terms of effectiveness, the performance of SVMs over the Yahoo! Directory is still far from satisfactory, which indicates that more substantial investigation is needed.

Feature selection for ranking

Ranking is a very important topic in information retrieval. While algorithms for learning ranking models have been intensively studied, this is not the case for feature selection, despite of its importance. The reality is that many feature selection methods used in classification are directly applied to ranking. We argue that because of the striking differences between ranking and classification, it is better to develop different feature selection methods for ranking. To this end, we propose a new feature selection method in this paper. Specifically, for each feature we use its value to rank the training instances, and define the ranking accuracy in terms of a performance measure or a loss function as the importance of the feature. We also define the correlation between the ranking results of two features as the similarity between them. Based on the definitions, we formulate the feature selection issue as an optimization problem, for which it is to find the features with maximum total importance scores and minimum total similarity scores. We also demonstrate how to solve the optimization problem in an efficient way. We have tested the effectiveness of our feature selection method on two information retrieval datasets and with two ranking models. Experimental results show that our method can outperform traditional feature selection methods for the ranking task.

Information Retrieval Technology

String similarity search and joins are primitive operations in database and information retrieval to address the poor data quality problem. Due to the high complexity of deletion neighborhoods, existing methods resort to hashing schemes to achieve reduction in space requirement of the index. However the introduced hash collisions need to be verified by the costly edit distance computation. In this paper, we focus on achieving a faster query speed with affordable memory consumptions. We propose a novel method that leverages the power of deletion neighborhoods and trie to answer the edit distance based string similarity query efficiently. We utilize the trie to share common prefixes of deletion neighborhoods and propose subtree merging optimization to reduce the index size. Then the index partition strategies are discussed and bit vector based verification method is proposed to speed up the query. The experimental results show that our method outperforms state-of-art methods on real dataset.

FRank: a ranking method with fidelity loss

Ranking problem is becoming important in many fields, especially in information retrieval (IR). Many machine learning techniques have been proposed for ranking problem, such as RankSVM, RankBoost, and RankNet. Among them, RankNet, which is based on a probabilistic ranking framework, is leading to promising results and has been applied to a commercial Web search engine. In this paper we conduct further study on the probabilistic ranking framework and provide a novel loss function named fidelity loss for measuring loss of ranking. The fidelity loss notonly inherits effective properties of the probabilistic ranking framework in RankNet, but possesses new properties that are helpful for ranking. This includes the fidelity loss obtaining zero for each document pair, and having a finite upper bound that is necessary for conducting query-level normalization. We also propose an algorithm named FRank based on a generalized additive model for the sake of minimizing the fedelity loss and learning an effective ranking function. We evaluated the proposed algorithm for two datasets: TREC dataset and real Web search dataset. The experimental results show that the proposed FRank algorithm outperforms other learning-based ranking methods on both conventional IR problem and Web search.

Query dependent ranking using K-nearest neighbor

Many ranking models have been proposed in information retrieval, and recently machine learning techniques have also been applied to ranking model construction. Most of the existing methods do not take into consideration the fact that significant differences exist between queries, and only resort to a single function in ranking of documents. In this paper, we argue that it is necessary to employ different ranking models for different queries and onduct what we call query-dependent ranking. As the first such attempt, we propose a K-Nearest Neighbor (KNN) method for query-dependent ranking. We first consider an online method which creates a ranking model for a given query by using the labeled neighbors of the query in the query feature space and then rank the documents with respect to the query using the created model. Next, we give two offline approximations of the method, which create the ranking models in advance to enhance the efficiency of ranking. And we prove a theory which indicates that the approximations are accurate in terms of difference in loss of prediction, if the learning algorithm used is stable with respect to minor changes in training examples. Our experimental results show that the proposed online and offline methods both outperform the baseline method of using a single ranking function.

Supervised rank aggregation

This paper is concerned with rank aggregation, the task of combining the ranking results of individual rankers at meta-search. Previously, rank aggregation was performed mainly by means of unsupervised learning. To further enhance ranking accuracies, we propose employing supervised learning to perform the task, using labeled data. We refer to the approach as Supervised Rank Aggregation. We set up a general framework for conducting Supervised Rank Aggregation, in which learning is formalized an optimization which minimizes disagreements between ranking results and the labeled data. As case study, we focus on Markov Chain based rank aggregation in this paper. The optimization for Markov Chain based methods is not a convex optimization problem, however, and thus is hard to solve. We prove that we can transform the optimization problem into that of Semidefinite Programming and solve it efficiently. Experimental results on meta-searches show that Supervised Rank Aggregation can significantly outperform existing unsupervised methods.

Web image clustering by consistent utilization of visual features and surrounding texts

Image clustering, an important technology for image processing, has been actively researched for a long period of time. Especially in recent years, with the explosive growth of the Web, image clustering has even been a critical technology to help users digest the large amount of online visual information. However, as far as we know, many previous works on image clustering only used either low-level visual features or surrounding texts, but rarely exploited these two kinds of information in the same framework. To tackle this problem, we proposed a novel method named consistent bipartite graph co-partitioning in this paper, which can cluster Web images based on the consistent fusion of the information contained in both low-level features and surrounding texts. In particular, we formulated it as a constrained multi-objective optimization problem, which can be efficiently solved by semi-definite programming (SDP). Experiments on a real-world Web image collection showed that our proposed method outperformed the methods only based on low-level features or surround texts.