Ranking is at the core of Information Retrieval.
Classic ranking optimization studies often treat ranking as a sorting problem with the assumption that the best performance of ranking would be achieved if we rank items according to their individual utility.
Accordingly, considerable ranking metrics have been developed and learning-to-rank algorithms that have been designed to optimize these simple performance metrics have been widely used in modern IR systems.
As applications evolve, however, people's need for information retrieval have shifted from simply retrieving relevant documents to more advanced information services that satisfy their complex working and entertainment needs.
Thus, more complicated and user-centric objectives such as user satisfaction and engagement have been adopted to evaluate modern IR systems today.
Those objectives, unfortunately, are difficult to be optimized under existing learning-to-rank frameworks as they are subject to great variance and complicated structures that cannot be explicitly explained or formulated with math equations like those simple performance metrics.
This leads to the following research question -- how to optimize result ranking for complex ranking metrics without knowing their internal structures?
To address this question, we conduct formal analysis on the limitation of existing ranking optimization techniques and describe three research tasks in \textit{Metric-agnostic Ranking Optimization}.
Through the discussion of potential solutions to these tasks, we hope to encourage more people to look into the problem of ranking optimization in complex search and recommendation scenarios.
Metric-agnostic Learning-to-Rank via Boosting and Rank Approximation
Dept. of Statistics \& Data Science University of Central Florida Orlando, USA；Department of CIS, SKL-IOTSC University of Macau Macao, China；School of Computing；AI Arizona State University Tempe, USA
The way pages are ranked in search results influences whether the users of search engines are exposed to more homogeneous, or rather to more diverse viewpoints. However, this viewpoint diversity is not trivial to assess. In this paper we use existing and novel ranking fairness metrics to evaluate viewpoint diversity in search result rankings. We conduct a controlled simulation study that shows how ranking fairness metrics can be used for viewpoint diversity, how their outcome should be interpreted, and which metric is most suitable depending on the situation. This paper lays out important grou…
The axiomatic analysis of IR evaluation metrics has contributed to a better understanding of their properties. Some works have modelled the effectiveness of retrieval measures with axioms that capture desirable properties on the set of ranked lists of documents. Recently, it has been shown that three of these axioms lead to some orderings. This work formally explores the metric properties of the set of rankings, endowed with these orderings. Based on lattice theory, the possible metrics and pseudo-metrics, defined on these structures, are determined. It is found that, when the relevant documen…
The study of IR evaluation metrics through axiomatic analysis enables a better understanding of their numerical properties. Some works have modelled the effectiveness of retrieval metrics with axioms that capture desirable properties on the set of rankings of documents. This paper formally explores the effect of these ranking axioms on the numerical values of some IR evaluation metrics. It focuses on the set of ranked lists of documents with multigrade relevance. The possible orderings in this set are derived from three commonly accepted ranking axioms on retrieval metrics; then, they are clas…
Ranked lists are frequently used by information retrieval (IR) systems to present results believed to be relevant to the users information need. Fairness is a relatively new but important aspect of these rankings to measure, joining a rich set of metrics that go beyond traditional accuracy or utility constructs to provide a more holistic understanding of IR system behavior. In the last few years, several metrics have been proposed to quantify the (un)fairness of rankings, particularly with respect to particular group(s) of content providers, but comparative analyses of these metrics -- particu…
Understanding why a model makes certain predictions is crucial when adapting it for real world decision making. LIME is a popular model-agnostic feature attribution method for the tasks of classification and regression. However, the task of learning to rank in information retrieval is more complex in comparison with either classification or regression. In this work, we extend LIME to propose Rank-LIME, a model-agnostic, local, post-hoc linear feature attribution method for the task of learning to rank that generates explanations for ranked lists. We employ novel correlation-based perturbations…
User behavior has been validated to be effective in revealing personalized preferences for commercial recommendations. However, few user-item interactions can be collected for new users, which results in a null space for their interests, i.e., the cold-start dilemma. In this paper, a two-tower framework, namely, the model-agnostic interest learning (MAIL) framework, is proposed to address the cold-start recommendation (CSR) problem for recommender systems. In MAIL, one unique tower is constructed to tackle the CSR from a zero-shot view, and the other tower focuses on the general ranking task. …
Recent work has proposed stochastic Plackett-Luce (PL) ranking models as a robust choice for optimizing relevance and fairness metrics. Unlike their deterministic counterparts that require heuristic optimization algorithms, PL models are fully differentiable. Theoretically, they can be used to optimize ranking metrics via stochastic gradient descent. However, in practice, the computation of the gradient is infeasible because it requires one to iterate over all possible permutations of items. Consequently, actual applications rely on approximating the gradient via sampling techniques. In this p…
State-of-the-art recommender system (RS) mostly rely on complex deep neural network (DNN) model structure, which makes it difficult to provide explanations along with RS decisions. Previous researchers have proved that providing explanations along with recommended items can help users make informed decisions and improve their trust towards the uninterpretable blackbox system. In model-agnostic explainable recommendation, system designers deploy a separate explanation model to take as input from the decision model, and generate explanations to meet the goal of persuasiveness. In this work, we e…
Learning to Rank (LTR) algorithms are usually evaluated using Information Retrieval metrics like Normalised Discounted Cumulative Gain (NDCG) or Mean Average Precision. As these metrics rely on sorting predicted items' scores (and thus, on items' ranks), their derivatives are either undefined or zero everywhere. This makes them unsuitable for gradient-based optimisation, which is the usual method of learning appropriate scoring functions. Commonly used LTR loss functions are only loosely related to the evaluation metrics, causing a mismatch between the optimisation objective and the evaluation…
Counterfactual Learning to Rank (LTR) methods optimize ranking systems using logged user interactions that contain interaction biases. Existing methods are only unbiased if users are presented with all relevant items in every ranking. There is currently no existing counterfactual unbiased LTR method for top-k rankings. We introduce a novel policy-aware counterfactual estimator for LTR metrics that can account for the effect of a stochastic logging policy. We prove that the policy-aware estimator is unbiased if every relevant item has a non-zero probability to appear in the top-k ranking. Our e…
This paper proposes a theoretical framework which models the information provided by retrieval systems in terms of Information Theory. The proposed framework allows to formalize: (i) system effectiveness as an information theoretic similarity between system outputs and human assessments, and (ii) ranking fusion as an information quantity measure. As a result, the proposed effectiveness metric improves popular metrics in terms of formal constraints. In addition, our empirical experiments suggest that it captures quality aspects from traditional metrics, while the reverse is not true. Our work a…
We introduce the concept of \emph{expected exposure} as the average attention ranked items receive from users over repeated samples of the same query. Furthermore, we advocate for the adoption of the principle of equal expected exposure: given a fixed information need, no item should receive more or less expected exposure than any other item of the same relevance grade. We argue that this principle is desirable for many retrieval objectives and scenarios, including topical diversity and fair ranking. Leveraging user models from existing retrieval metrics, we propose a general evaluation method…
Direct optimization of IR metrics has often been adopted as an approach to devise and develop ranking-based recommender systems. Most methods following this approach aim at optimizing the same metric being used for evaluation, under the assumption that this will lead to the best performance. A number of studies of this practice bring this assumption, however, into question. In this paper, we dig deeper into this issue in order to learn more about the effects of the choice of the metric to optimize on the performance of a ranking-based recommender system. We present an extensive experimental st…
A recommender system generates personalized recommendations for a user by computing the preference score of items, sorting the items according to the score, and filtering top-K items with high scores. While sorting and ranking items are integral for this recommendation procedure, it is nontrivial to incorporate them in the process of end-to-end model training since sorting is nondifferentiable and hard to optimize with gradient descent. This incurs the inconsistency issue between existing learning objectives and ranking metrics of recommenders. In this work, we present DRM (differentiable rank…
Ranking evaluation metrics are a fundamental element of design and improvement efforts in information retrieval. We observe that most popular metrics disregard information portrayed in the scores used to derive rankings, when available. This may pose a numerical scaling problem, causing an under- or over-estimation of the evaluation depending on the degree of divergence between the scores of ranked items. The purpose of this work is to propose a principled way of quantifying multi-graded relevance judgments of items and enable a more accurate penalization of ordering errors in rankings. We pro…
L3S Research Center（L3S研究中心）；Leibniz Universität Hannover（汉诺威莱布尼茨大学）
Learning-to-rank (LTR) is a class of supervised learning techniques that apply to ranking problems dealing with a large number of features. The popularity and widespread application of LTR models in prioritizing information in a variety of domains makes their scrutability vital in today's landscape of fair and transparent learning systems. However, limited work exists that deals with interpreting the decisions of learning systems that output rankings. In this paper we propose a model agnostic local explanation method that seeks to identify a small subset of input features as explanation to a r…
Information retrieval (IR) systems traditionally aim to maximize metrics built on rankings, such as precision or NDCG. However, the non-differentiability of the ranking operation prevents direct optimization of such metrics in state-of-the-art neural IR models, which rely entirely on the ability to compute meaningful gradients. To address this shortcoming, we propose SmoothI, a smooth approximation of rank indicators that serves as a basic building block to devise differentiable approximations of IR metrics. We further provide theoretical guarantees on SmoothI and derived approximations, showi…
University College London（伦敦大学学院）；Microsoft Research（微软研究院）；Max Planck Institute for Security and Privacy（马克斯·普朗克安全与隐私研究所）；Boise State University（博伊西州立大学）；Spotify Research（Spotify研究院）
There is increasing attention to evaluating the fairness of search system ranking decisions. These metrics often consider the membership of items to particular groups, often identified using protected attributes such as gender or ethnicity. To date, these metrics typically assume the availability and completeness of protected attribute labels of items. However, the protected attributes of individuals are rarely present, limiting the application of fair ranking metrics in large scale systems. In order to address this problem, we propose a sampling strategy and estimation technique for four fair…
Evaluation metrics are an essential part of a ranking system, and in the past many evaluation metrics have been proposed in information retrieval and Web search. Discounted Cumulated Gains (DCG) has emerged as one of the evaluation metrics widely adopted for evaluating the performance of ranking functions used in Web search. However, the two sets of parameters, gain values and discount factors, used in DCG are determined in a rather ad-hoc way. In this paper we first show that DCG is generally not coherent, meaning that comparing the performance of ranking functions using DCG very much depends…
Despite the increasing use of citation-based metrics for research evaluation purposes, we do not know yet which metrics best deliver on their promise to gauge the significance of a scientific paper or a patent. We assess 17 network-based metrics by their ability to identify milestone papers and patents in three large citation datasets. We find that traditional information-retrieval evaluation metrics are strongly affected by the interplay between the age distribution of the milestone items and age biases of the evaluated metrics. Outcomes of these metrics are therefore not representative of th…
University of Mannheim（曼海姆大学）；L3S Research Center（L3S研究中心）；Leibniz University Hannover（汉诺威莱布尼茨大学）；GESIS - Leibniz Institute for the Social Sciences（GESIS - 莱布尼茨社会科学研究所）；Complexity Science Hub（复杂科学中心）
Computer Vision Center（计算机视觉中心）；Universitat Autònoma de Barcelona（巴塞罗那自治大学）
In this paper, we explore and evaluate the use of ranking-based objective functions for learning simultaneously a word string and a word image encoder. We consider retrieval frameworks in which the user expects a retrieval list ranked according to a defined relevance score. In the context of a word spotting problem, the relevance score has been set according to the string edit distance from the query string. We experimentally demonstrate the competitive performance of the proposed model on query-by-string word spotting for both, handwritten and real scene word images. We also provide the resul…
Learning to rank is an effective recommendation approach since its introduction around 2010. Famous algorithms such as Bayesian Personalized Ranking and Collaborative Less is More Filtering have left deep impact in both academia and industry. However, most learning to rank approaches focus on improving technical accuracy metrics such as AUC, MRR and NDCG. Other evaluation metrics of recommender systems like fairness have been largely overlooked until in recent years. In this paper, we propose a new learning to rank algorithm named Pareto Pairwise Ranking. We are inspired by the idea of Bayesia…
Towards Fair Rankings: Leveraging LLMs for Gender Bias Detection and Measurement
Università della Svizzera italiana \& University of Amsterdam；University of Amsterdam The Netherland；Unviersity of Amsterdam The Netherland；University of Amsterdam
Nanyang Technological University（南洋理工大学）；Institute for Infocomm Research（信息通信研究院）
This paper proposes a new neural architecture for collaborative ranking with implicit feedback. Our model, LRML (\textit{Latent Relational Metric Learning}) is a novel metric learning approach for recommendation. More specifically, instead of simple push-pull mechanisms between user and item pairs, we propose to learn latent relations that describe each user item interaction. This helps to alleviate the potential geometric inflexibility of existing metric learing approaches. This enables not only better performance but also a greater extent of modeling capability, allowing our model to scale t…
Static rankings of papers play a key role in the academic search setting. Many features are commonly used in the literature to produce such rankings, some examples are citation-based metrics, distinct applications of PageRank, among others. More recently, learning to rank techniques have been successfully applied to combine sets of features producing effective results. In this work, we propose the metric S-RCR, which is a simplified version of a metric called Relative Citation Ratio --- both based on the idea of a co-citation network. When compared to the classical version, our simplification …
University of Pittsburgh（匹兹堡大学）；Yahoo! Research（雅虎研究院）；Airbnb（爱彼迎）；University of California, Los Angeles（加利福尼亚大学洛杉矶分校）
The content ranking problem in a social news website, is typically a function that maximizes a scalar metric of interest like dwell-time. However, like in most real-world applications we are interested in more than one metric---for instance simultaneously maximizing click-through rate, monetization metrics, dwell-time---and also satisfy the traffic requirements promised to different publishers. All this needs to be done on online data and under the settings where the objective function and the constraints can dynamically change; this could happen if for instance new publishers are added, some …
Connections among entities are everywhere. From social media interactions to web page hyperlinks, networks are frequently used to represent such complex systems. Node ranking is a fundamental task that provides the strategy to identify central entities according to multiple criteria. Popular node ranking metrics include degree, closeness or betweenness centralities, as well as HITS authority or PageRank. In this work, we propose a novel node ranking metric, where we combine PageRank and the idea of node fatigue, in order to model a random explorer who wants to optimize coverage - it gets fatig…
The task of item recommendation requires ranking a large catalogue of items given a context. Item recommendation algorithms are evaluated using ranking metrics that depend on the positions of relevant items. To speed up the computation of metrics, recent work often uses sampled metrics where only a smaller set of random items and the relevant items are ranked. This paper investigates sampled metrics in more detail and shows that sampled metrics are inconsistent with their exact version. Sampled metrics do not persist relative statements, e.g., 'algorithm A is better than B', not even in expect…
We extend the recently introduced theory of Lovasz-Bregman (LB) divergences (Iyer & Bilmes, 2012) in several ways. We show that they represent a distortion between a 'score' and an 'ordering', thus providing a new view of rank aggregation and order based clustering with interesting connections to web ranking. We show how the LB divergences have a number of properties akin to many permutation based metrics, and in fact have as special cases forms very similar to the Kendall-$τ$ metric. We also show how the LB divergences subsume a number of commonly used ranking measures in information retrieva…
Today, ranking systems in universities have been considered by the academic community, and there is a tight competition between world universities to achieve higher ranks. In the meantime, the ranking of university websites is also in the spotlight, and the Webometric research center announces the ranks of university websites twice a year. Examining university rankings indicators and the Webometric ranks of the university indicates that some of these indicators, directly and indirectly, affect each other. On the other hand, a preliminary study of Webometric indicators shows that some Search En…
Johns Hopkins University（约翰斯·霍普金斯大学）；University of Cambridge（剑桥大学）；Microsoft Research（微软研究院）
Learning to rank -- producing a ranked list of items specific to a query and with respect to a set of supervisory items -- is a problem of general interest. The setting we consider is one in which no analytic description of what constitutes a good ranking is available. Instead, we have a collection of representations and supervisory information consisting of a (target item, interesting items set) pair. We demonstrate analytically, in simulation, and in real data examples that learning to rank via combining representations using an integer linear program is effective when the supervision is as …
University of Colorado, Boulder（科罗拉多大学博尔德分校）；DePaul University（德保罗大学）；The Chinese University of Hong Kong（香港中文大学）
As recommender systems have become more widespread and moved into areas with greater social impact, such as employment and housing, researchers have begun to seek ways to ensure fairness in the results that such systems produce. This work has primarily focused on developing recommendation approaches in which fairness metrics are jointly optimized along with recommendation accuracy. However, the previous work had largely ignored how individual preferences may limit the ability of an algorithm to produce fair recommendations. Furthermore, with few exceptions, researchers have only considered sce…
Microsoft Research, Redmond（微软研究院，雷德蒙德）；Duke University（杜克大学）；Brown University（布朗大学）
Collaborative filtering is widely used in modern recommender systems. Recent research shows that variational autoencoders (VAEs) yield state-of-the-art performance by integrating flexible representations from deep neural networks into latent variable models, mitigating limitations of traditional linear factor models. VAEs are typically trained by maximizing the likelihood (MLE) of users interacting with ground-truth items. While simple and often effective, MLE-based training does not directly maximize the recommendation-quality metrics one typically cares about, such as top-N ranking. In this …
Recommender systems are one of the most pervasive applications of machine learning in industry, with many services using them to match users to products or information. As such it is important to ask: what are the possible fairness risks, how can we quantify them, and how should we address them? In this paper we offer a set of novel metrics for evaluating algorithmic fairness concerns in recommender systems. In particular we show how measuring fairness based on pairwise comparisons from randomized experiments provides a tractable means to reason about fairness in rankings from recommender syst…
Google Research（谷歌研究院）；National Institutes of Health（美国国立卫生研究院）
Listwise learning-to-rank methods form a powerful class of ranking algorithms that are widely adopted in applications such as information retrieval. These algorithms learn to rank a set of items by optimizing a loss that is a function of the entire set -- as a surrogate to a typically non-differentiable ranking metric. Despite their empirical success, existing listwise methods are based on heuristics and remain theoretically ill-understood. In particular, none of the empirically successful loss functions are related to ranking metrics. In this work, we propose a cross entropy-based learning-to…
Amazon（亚马逊）；University of California, Los Angeles（加利福尼亚大学洛杉矶分校）；Facebook AI Research（Facebook人工智能研究院）；Stanford University（斯坦福大学）
A key challenge with machine learning approaches for ranking is the gap between the performance metrics of interest and the surrogate loss functions that can be optimized with gradient-based methods. This gap arises because ranking metrics typically involve a sorting operation which is not differentiable w.r.t. the model parameters. Prior works have proposed surrogates that are loosely related to ranking metrics or simple smoothed versions thereof, and often fail to scale to real-world applications. We propose PiRank, a new class of differentiable surrogates for ranking, which employ a continu…
Several tasks in machine learning are evaluated using non-differentiable metrics such as mean average precision or Spearman correlation. However, their non-differentiability prevents from using them as objective functions in a learning framework. Surrogate and relaxation methods exist but tend to be specific to a given metric. In the present work, we introduce a new method to learn approximations of such non-differentiable objective functions. Our approach is based on a deep architecture that approximates the sorting of arbitrary sets of scores. It is trained virtually for free using synthetic…
