Top-k Ranking Research Articles

Tri-Branch Convolutional Neural Networks for Top-k Focused Academic Performance Prediction.

Academic performance prediction aims to leverage student-related information to predict their future academic outcomes, which is beneficial to numerous educational applications, such as personalized teaching and academic early warning. In this article, we reveal the students' behavior trajectories by mining campus smartcard records, and capture the characteristics inherent in trajectories for academic performance prediction. Particularly, we carefully design a tri-branch convolutional neural network (CNN) architecture, which is equipped with rowwise, columnwise, and depthwise convolutions and attention operations, to effectively capture the persistence, regularity, and temporal distribution of student behavior in an end-to-end manner, respectively. However, different from existing works mainly targeting at improving the prediction performance for the whole students, we propose to cast academic performance prediction as a top-k ranking problem, and introduce a top-k focused loss to ensure the accuracy of identifying academically at-risk students. Extensive experiments were carried out on a large-scale real-world dataset, and we show that our approach substantially outperforms recently proposed methods for academic performance prediction. For the sake of reproducibility, our codes have been released at https://github.com/ZongJ1111/Academic-Performance-Prediction.

Attention guided domain alignment for conditional face image generation

Despite the great success of Generative Adversarial Network (GAN) in face image generation, it is still a challenge to generate a faithful yet high-fidelity face image given an exemplar image and a conditional input from a distinct domain (e.g., a semantic segmentation mask or an edge map). Existing methods learn image-level features to align distinct domains in an intermediate domain, which ignore the spatial relationship of the facial semantic parts and therefore lead to semantic mismatching problems. In addition, it is computationally expensive to establish dense correspondences in the intermediate domain, especially for high-resolution face images. To address these problems, we propose a novel attention guided domain alignment method for conditional face image generation, which aligns two domains directly under the guidance of the local attention learned from semantically similar face parts. In particular, we assign a dedicated index for each feature block and adopt a top-k ranking operation to correspond block-wise features in two distinct domains, which exploits the spatial relationship of facial parts and preserves the texture structure during alignment. The local attention is then learned from the retrieved blocks, which reduces computation complexity substantially and therefore enables to build alignment in a high resolution. The aligned features are finally fused with adaptive weights learned from their long-range correlation coefficients, which capture the semantic coherence of the style features between the two domains. Extensive experimental results on the CelebAMask-HQ dataset demonstrate that the proposed method is superior to the state-of-the-art methods.

Why Not Yet: Fixing a Top-k Ranking that is Not Fair to Individuals

This work considers why-not questions in the context of top-k queries and score-based ranking functions. Following the popular linear scalarization approach for multi-objective optimization, we study rankings based on the weighted sum of multiple scores. A given weight choice may be controversial or perceived as unfair to certain individuals or organizations, triggering the question why some entity of interest has not yet shown up in the top-k. We introduce various notions of such why-not-yet queries and formally define them as satisfiability or optimization problems, whose goal is to propose alternative ranking functions that address the placement of the entities of interest. While some why-not-yet problems have linear constraints, others require quantifiers, disjunction, and negation. We propose several optimizations, ranging from a monotonic-core construction that approximates the complex constraints with a conjunction of linear ones, to various techniques that let the user control the tradeoff between running time and approximation quality. Experiments with real and synthetic data demonstrate the practicality and scalability of our technique, showing its superiority compared to the state of the art (SOA).

Collaborative Social Metric Learning in Trust Network for Recommender Systems

In this study, a novel top-K ranking recommendation method called collaborative social metric learning (CSML) is proposed, which implements a trust network that provides both user-item and user-user interactions in simple structure. Most existing recommender systems adopting trust networks focus on item ratings, but this does not always guarantee optimal top-K ranking prediction. Conventional direct ranking systems in trust networks are based on sub-optimal correlation approaches that do not consider item-item relations. The proposed CSML algorithm utilizes the metric learning method to directly predict the top-K items in a trust network. A new triplet loss is further proposed, called socio-centric loss, which represents user-user interactions to fully exploit the information contained in a trust network, as an addition to the two commonly used triplet losses in metric learning for recommender systems, which consider user-item and item-item relations. Experimental results demonstrate that the proposed CSML outperformed existing recommender systems for real-world trust network data.

Partial recovery for top-k ranking: Optimality of MLE and SubOptimality of the spectral method

Given partially observed pairwise comparison data generated by the Bradley–Terry–Luce (BTL) model, we study the problem of top-k ranking. That is, to optimally identify the set of top-k players. We derive the minimax rate with respect to a normalized Hamming loss. This provides the first result in the literature that characterizes the partial recovery error in terms of the proportion of mistakes for top-k ranking. We also derive the optimal signal to noise ratio condition for the exact recovery of the top-k set. The maximum likelihood estimator (MLE) is shown to achieve both optimal partial recovery and optimal exact recovery. On the other hand, we show another popular algorithm, the spectral method, is in general suboptimal. Our results complement the recent work (Ann. Statist. 47 (2019) 2204–2235) that shows both the MLE and the spectral method achieve the optimal sample complexity for exact recovery. It turns out the leading constants of the sample complexity are different for the two algorithms. Another contribution that may be of independent interest is the analysis of the MLE without any penalty or regularization for the BTL model. This closes an important gap between theory and practice in the literature of ranking.

DualGCN: An Aspect-Aware Dual Graph Convolutional Network for review-based recommender

Recently, a variety of review-based recommendation systems that incorporate the valuable information extracted from user-generated textual reviews into user and item modeling have been proposed. However, the existing recommendations normally model reviews at the sentence level. They ignore the modeling of aspect words in reviews, which fails to capture user preferences and item attributes in a fine-grained way. In addition, few studies consider constructing user–item interaction based on review information extracted from the aspect level. In this paper, we are motivated to propose an Aspect-Aware Dual Graph Convolutional Network (DualGCN). Specifically, we first design an Aspect-GCN layer to model the message diffusion of an aspect graph constructed from reviews, capturing the overall description of an aspect in all reviews. We then propose a UI-GCN layer to model a user’s fine-grained preferences toward interacted items at the aspect level. Finally, we adopt a Factorization Machine model to accomplish the recommendation task. The experimental results demonstrate that our model significantly outperforms the related approaches w.r.t. the accuracy of both rating prediction and top-K ranking on Amazon and Yelp datasets.

Optimizing Partial Area Under the Top-k Curve: Theory and Practice.

Top- k error has become a popular metric for large-scale classification benchmarks due to the inevitable semantic ambiguity among classes. Existing literature on top- k optimization generally focuses on the optimization method of the top- k objective, while ignoring the limitations of the metric itself. In this paper, we point out that the top- k objective lacks enough discrimination such that the induced predictions may give a totally irrelevant label a top rank. To fix this issue, we develop a novel metric named partial Area Under the top- k Curve (AUTKC). Theoretical analysis shows that AUTKC has a better discrimination ability, and its Bayes optimal score function could give a correct top- K ranking with respect to the conditional probability. This shows that AUTKC does not allow irrelevant labels to appear in the top list. Furthermore, we present an empirical surrogate risk minimization framework to optimize the proposed metric. Theoretically, we present (1) a sufficient condition for Fisher consistency of the Bayes optimal score function; (2) a generalization upper bound which is insensitive to the number of classes under a simple hyperparameter setting. Finally, the experimental results on four benchmark datasets validate the effectiveness of our proposed framework.

An Enhanced Re-Ranking Model for Person Re-Identification

Presently, Person Re-IDentification (PRe-ID) acts as a vital part of real time video surveillance to ensure the rising need for public safety. Resolving the PRe-ID problem includes the process of matching observations of persons among distinct camera views. Earlier models consider PRe-ID as a unique object retrieval issue and determine the retrieval results mainly based on the unidirectional matching among the probe and gallery images. But the accurate matching might not be present in the top-k ranking results owing to the appearance modifications caused by the difference in illumination, pose, viewpoint, and occlusion. For addressing these issues, a new Hyper-parameter Optimized Deep Learning (DL) approach with Expanded Neighborhood Distance Reranking (HPO-DLDN) model is proposed for PRe-ID. The proposed HPO-DLDN involves different processes for PRe-ID, such as feature extraction, similarity measurement, and feature re-ranking. The HPO-DLDN model uses a Adam optimizer with Densely Connected Convolutional Networks (DenseNet169) model as a feature extractor. Additionally, Euclidean distance-based similarity measurement is employed to determine the resemblance between the probe and gallery images. Finally, the HPO-DLDN model incorporated ENDR model to re-rank the outcome of the person-reidentification along with Mahalanobis distance. An extensive experimental analysis is carried out on CUHK01 benchmark dataset and the obtained results verified the effective performance of the HPO-DLDN model in different aspects.

Periodical Top-kRanking Query in Wireless Sensor Networks Exploiting Wake-Up Receivers

This paper proposes an energy–efficient wake–up control for periodical top– <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${k}$ </tex-math></inline-formula> query in wireless sensor networks employing wake–up receivers. The wake–up procedure is designed in such a way that it only activates a subset of nodes that contribute to grasping the information on the top– <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${k}$ </tex-math></inline-formula> ranking set with high accuracy, which prevents unnecessary wake–up of irrelevant nodes. The proposed scheme exploits the temporal correlation of observed data with a transmission suppression control executed according to the previous top– <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${k}$ </tex-math></inline-formula> set, which is notified from the sink to wake–up receivers by a designated signaling. Furthermore, the validations on the estimated top– <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${k}$ </tex-math></inline-formula> set are conducted by exploiting two different types of wake–up control: identity–based wake–up and content–based wake–up. With computer simulations, we show that our proposed scheme can finely control the trade–off between the accuracy on top– <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${k}$ </tex-math></inline-formula> set and energy efficiency, and it can achieve high energy saving of sensor nodes in comparison to the other wake–up control designed for the top– <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${k}$ </tex-math></inline-formula> query. We also investigate the practicality of the proposed wake–up control by implementing its operations into a prototype of sensor node with wake–up receiver. Furthermore, we analyze the computational complexity of our proposed wake–up control.

Fair Top-k Ranking with multiple protected groups

Ranking items or people is a fundamental operation at the basis of several processes and services, not all of them happening online. Ranking is required for different tasks, including search, personalization, recommendation, and filtering. While traditionally ranking has been aimed solely at maximizing some global utility function, recently the awareness of potential discrimination for some of the elements to rank, has captured the attention of researchers, which have thus started devising ranking systems which are non-discriminatory or fair for the items being ranked. So far, researchers have mostly focused on group fairness, which is usually expressed in the form of constraints on the fraction of elements from some protected groups that should be included in the top-k positions, for any relevant k. These constraints are needed in order to correct implicit societal biases existing in the input data and reflected in the relevance or fitness score computed.In this article, we tackle the problem of selecting a subset of k individuals from a pool of n≫k candidates, maximizing global utility (i.e., selecting the “best” candidates) while respecting given group-fairness criteria. In particular, to tackle this Fair Top-kRanking problem, we adopt a ranked group-fairness definition which extends the standard notion of group fairness based on protected groups, by ensuring that the proportion of protected candidates in every prefix of the top-k ranking remains statistically above, or indistinguishable from, a given minimum threshold. Our notion of utility requires, intuitively, that every individual included in the top-k should be more qualified than every candidate not included; and that for every pair of candidates in the top-k, the more qualified candidate should be ranked above.The main contribution of this paper is an algorithm for producing a fair top-k ranking that can be used when more than one protected group is present, which means that a statistical test based on a multinomial distribution needs to be used instead of one for a binomial distribution, as the original FA*IR algorithms does. This poses important technical challenges and increases both the space and time complexity of the re-ranking algorithm. Our experimental assessment on real-world datasets shows that our approach yields small distortions with respect to rankings that maximize utility without considering our fairness criteria.

Top-k Ranking Bayesian Optimization

This paper presents a novel approach to top-k ranking Bayesian optimization (top-k ranking BO) which is a practical and significant generalization of preferential BO to handle top-k ranking and tie/indifference observations. We first design a surrogate model that is not only capable of catering to the above observations, but is also supported by a classic random utility model. Another equally important contribution is the introduction of the first information-theoretic acquisition function in BO with preferential observation called multinomial predictive entropy search (MPES) which is flexible in handling these observations and optimized for all inputs of a query jointly. MPES possesses superior performance compared with existing acquisition functions that select the inputs of a query one at a time greedily. We empirically evaluate the performance of MPES using several synthetic benchmark functions, CIFAR-10 dataset, and SUSHI preference dataset.

Research on location and similar comments in point-of-interest recommendation system for users

By combining deep learning algorithms and potential probabilistic topic models, this paper proposes a new modeling strategy for individual users and constructs an effective point of interest recommendation system. First, Doc2vec is used to vectorize the user’s comments, and the Bow model is used to match the similarity of the points of interest, and then the potential probability model is used to cluster the positions of the points of interest. Finally, relying on the user’s historical comment data as a basis, the Top-K ranking recommendation scheme is obtained. The experimental results show that for large-scale location-based social network data sets, the algorithm in this paper has an ideal final effect on users’ personalized recommendations. The recommendation model for user location information and similar comment points of interest proposed in this paper is more subjectively and objectively in line with user behavior needs. On the premise of ensuring the real-time performance of recommendation, the model proposed in this paper effectively improves the accuracy of recommendation compared with traditional point-of-interest recommendation algorithms.

Deep Top-$k$ Ranking for Image–Sentence Matching

Image-sentence matching is a challenging task for the heterogeneity-gap between different modalities. Ranking-based methods have achieved excellent performance in this task in past decades. Given an image query, these methods typically assume that the correct matched image-sentence pair must rank before all other mismatched ones. However, this assumption may be too strict and prone to the overfitting problem, especially when some sentences in a massive database are similar and confusable with one another. In this paper, we relax the traditional ranking loss and propose a novel deep multi-modal network with a top-k ranking loss to mitigate the data ambiguity problem. With this strategy, query results will not be penalized unless the index of ground truth is outside the range of top-k query results. Considering the non-smoothness and non-convexity of the initial top-k ranking loss, we exploit a tight convex upper bound to approximate the loss and then utilize the traditional back-propagation algorithm to optimize the deep multi-modal network. Finally, we apply the method on three benchmark datasets, namely, Flickr8k, Flickr30k, and MSCOCO. Empirical results on metrics R@K (K = 1, 5, 10) show that our method achieves comparable performance in comparison to state-of-the-art methods.

Optimizing target selection complexity of a recommendation system by skyline query and multi-criteria decision analysis

Various factors related to user consideration cause a target selection problem that may lead users to receive unexpected or confusing results. Traditionally, the recommendation system is constructed to help the user filter out unrelated targets and recommend targets that may be of interest to the user. However, the complexity of target selection requires a more advanced decision-making analysis to offer support. Determining how to optimize the target selection complexity of a recommendation system has become a critical challenge. This study proposes a novel approach using skyline query and multi-criteria decision analysis to recommend Top-k targets for user selection. Skyline query domination reduces the complexity of target selection by filtering out non-dominant candidates and keeping the dominant candidates for multi-criteria decision analysis. After the skyline query processing, the multi-criteria decision analysis is optimized, producing a Top-k ranking order of the candidate targets. The experiment illustrates an empirical case study to verify the effectiveness of the proposed approach. The contribution is optimizing the target selecting complexity of the recommendation system to solve the target selection problem.

An Inductive Content-Augmented Network Embedding Model for Edge Artificial Intelligence

Real-time data processing applications demand dynamic resource provisioning and efficient service discovery, which is particularly challenging in resource-constraint edge computing environments. Network embedding techniques can potentially aid effective resource discovery services in edge environments, by achieving a proximity-preserving representation of the network resources. Most of the existing techniques of network embedding fail to capture accurate proximity information among the network nodes and further lack exploiting information beyond the second-order neighbourhood. This paper leverages artificial intelligence for network representation and proposes a deep learning model, named inductive content augmented network embedding (ICANE), which integrates the network structure and resource content attributes into a feature vector. Secondly, a hierarchical aggregation approach is introduced to explicitly learn the network representation through sampling the nodes and aggregating features from the higher-order neighbourhood. A semantic proximity search model is then designed to generate the top-k ranking of relevant nodes using the learned network representation. Experiments conducted on real-world datasets demonstrate the superiority of the proposed model over the existing popular methods in terms of resource discovery and the query resolving performance.

SPECTRAL METHOD AND REGULARIZED MLE ARE BOTH OPTIMAL FOR TOP-K RANKING.

This paper is concerned with the problem of top-K ranking from pairwise comparisons. Given a collection of n items and a few pairwise comparisons across them, one wishes to identify the set of K items that receive the highest ranks. To tackle this problem, we adopt the logistic parametric model - the Bradley-Terry-Luce model, where each item is assigned a latent preference score, and where the outcome of each pairwise comparison depends solely on the relative scores of the two items involved. Recent works have made significant progress towards characterizing the performance (e.g. the mean square error for estimating the scores) of several classical methods, including the spectral method and the maximum likelihood estimator (MLE). However, where they stand regarding top-K ranking remains unsettled. We demonstrate that under a natural random sampling model, the spectral method alone, or the regularized MLE alone, is minimax optimal in terms of the sample complexity - the number of paired comparisons needed to ensure exact top-K identification, for the fixed dynamic range regime. This is accomplished via optimal control of the entrywise error of the score estimates. We complement our theoretical studies by numerical experiments, confirming that both methods yield low entrywise errors for estimating the underlying scores. Our theory is established via a novel leave-one-out trick, which proves effective for analyzing both iterative and non-iterative procedures. Along the way, we derive an elementary eigenvector perturbation bound for probability transition matrices, which parallels the Davis-Kahan theorem for symmetric matrices. This also allows us to close the gap between the error upper bound for the spectral method and the minimax lower limit.

Time-homogeneous top-K ranking using tensor decompositions

ABSTRACT Given N items, all having positive latent strengths, top-K ranking problem aims to identify the K items () receiving the highest ranks based on partially revealed comparisons among the items. This problem has been widely studied in the case for which comparisons are performed independently. However, identifying the top-K rankings becomes a more intricate task when comparisons per se are performed along some temporal dimension. In this paper, we investigate potential impacts of temporality on sequences of top-K items and propose a time-homogeneous ranking scheme. Our framework relies mainly on tensor decompositions, rank centrality, and an innovative continuous extension of the Bradley–Terry–Luce (BTL) model. The proposed continuous BTL model extends the win/loss nature of the logistic model to a continuous setting, further reflecting preference degrees that may exist among the compared items. Our computations, which pertain to the analysis of S&P500 data from January 2008 to December 2017, confirm that the proposed top-K ranking scheme is an effective approach to optimize cardinality-constrained portfolios which involve large volumes of noisy and incomplete data.

An efficient top-k ranking method for service selection based on ε-ADMOPSO algorithm

One of the main concerns in rank aggregation tasks for metasearch service is how to retrieve and aggregate the large-scale candidate search results efficiently. Much work has been done to implement metasearch service engines with different rank aggregation algorithms. However, the performance of these metasearch engines can hardly be improved. In this paper, we transform the top-k ranking task into a multi-objective programming problem when user preferences are considered along with user queries. We build an improved discrete multi-objective programming model to make the aggregate rankings satisfy user queries and user preferences both, and then propose a user preferences-based rank aggregation algorithm accordingly. Based on discrete particle swarm optimization algorithm, we improve the encoding scheme, the initialization methods, the position and velocity definition, the integrating updating process, the turbulence operator, and the external archive updating and leader selection strategy to make sure the candidate results that fit the user’s preferences can be located quickly and accurately in a large-scale discrete solution space. We have our proposed algorithm tested on three different benchmark datasets: a public dataset, the real-world datasets and the synthetic simulation datasets. The experimental results demonstrate the efficacy and convergence efficiency of the proposed algorithm over the baseline rank aggregation methods especially when dealing with large amount of candidate results. And when the set of candidate results is of normal size, the proposed algorithm is proved to perform not worse than the baseline methods.

TIIREC: A tensor approach for tag-driven item recommendation with sparse user generated content

In recent years, tagging system has become a building block o summarize the content of items for further functions like retrieval or personalized recommendation in various web applications. One nontrivial requirement is to precisely deliver a list of suitable items when users interact with the systems via inputing a specific tag (i.e. a query term). Different from traditional recommender systems, we need deal with a collaborative retrieval (CR) problem, where both characteristics of retrieval and recommendation should be considered to model a ternary relationship involved with query× user× item. Recently, several works are proposed to study CR task from users’ perspective. However, they miss a significant challenge raising from the sparse content of items. In this work, we argue that items will suffer from the sparsity problem more severely than users, since items are usually observed with fewer features to support a feature-based or content-based algorithm. To tackle this problem, we aim to sufficiently explore the sophisticated relationship of each query× user× item triple from items’ perspective. By integrating item-based collaborative information for this joint task, we present an alternative factorized model that could better evaluate the ranks of those items with sparse information for the given query–user pair. In addition, we suggest to employ a recently proposed Bayesian Personalized Ranking (BPR) algorithm to optimize latent collaborative retrieval problem from pairwise learning perspective. The experimental results on two real-world datasets, (i.e. Last.fm, Yelp), verified the efficiency and effectiveness of our proposed approach at top-k ranking metric.

Improving top-K recommendation with truster and trustee relationship in user trust network

Due to the data sparsity problem, social network information is often additionally used to improve the performance of recommender systems. While most existing works exploit social information to reduce the rating prediction error, e.g., RMSE, a few had aimed to improve the top-k ranking prediction accuracy. This paper proposes a novel top-k ranking oriented recommendation method, TRecSo, which incorporates social information into recommendation by modeling two different roles of users as trusters and trustees while considering the structural information of the network. Empirical studies on real-world datasets demonstrate that TRecSo leads to a remarkable improvement compared with previous methods in top-k recommendation.