Negative Can Be Positive: Signed Graph Neural Networks for Recommendation

Most of the existing GNN-based recommender system models focus on learning users’ personalized preferences from these (explicit/implicit) positive feedback to achieve personalized recommendations. However, in the real-world recommender system, the users’ feedback behavior also includes negative feedback behavior (e.g., click dislike button), which also reflects users’ personalized preferences. How to utilize negative feedback is a challenging research problem. In this paper, we first qualitatively and quantitatively analyze the three kinds of negative feedback that widely existed in real-world recommender systems and investigate the role of negative feedback in recommender systems. We found that it is different from what we expected — not all negative items are ranked low, and some negative items are even ranked high in the overall items. Then, we propose a novel Signed Graph Neural Network Recommendation model (SiGRec) to encode the users’ negative feedback behavior. Our SiGRec can learn positive and negative embeddings of users and items via positive and negative graph neural network encoders, respectively. Besides, we also define a new Sign Cosine (SiC) loss function to adaptively mine the information of negative feedback for different types of negative feedback. Extensive experiments on four datasets demonstrate the proposed model outperforms several existing models. Specifically, on the Zhihu dataset, SiGRec outperforms the unsigned GNN model (i.e., LightGCN), 27.58% 29.81%, and 31.21% in P@20, R@20, and nDCG@20, respectively. We hope our work can open the door to further exploring the negative feedback in recommendations.     

Information retrieval with geographical references. Relevant documents filtering vs. query expansion

This is a thorough analysis of two techniques applied to Geographic Information Retrieval (GIR). Previous studies have researched the application of query expansion to improve the selection process of information retrieval systems. This paper emphasizes the effectiveness of the filtering of relevant documents applied to a GIR system, instead of query expansion. Based on the CLEF (Cross Language Evaluation Forum) framework available, several experiments have been run. Some based on query expansion, some on the filtering of relevant documents. The results show that filtering works better in a GIR environment, because relevant documents are not reordered in the final list.     

信息内容特征对移动医疗APP用户推荐行为的影响及作用路径分析

[目的/意义]旨在探究信息内容特征对移动医疗APP用户推荐行为的影响和作用路径，提炼出对运营和管理具有积极意义的意见和方法。[方法/过程]以精细加工可能性模型为理论框架，通过整合信息系统成功模型和口碑营销经典理论的相关因素，构建出"中心信息内容特征——自我推荐"以及"边缘信息内容特征——从众推荐"两条作用路径，并聚焦感知需求的调节作用提出假设并构建研究模型。实证研究采用"日志追踪实验+调查问卷"的方式收集来自移动医疗APP用户的288份数据，利用SmartPLS2.0对模型进行检验。[结果/结论]信息质量、服务质量、系统质量正向影响自我推荐意愿；APP认知声誉、医生认知声誉、医院认知声誉正向影响从众推荐意愿；感知需求正向调节自我推荐意愿对推荐行为的影响，而从众推荐意愿对推荐行为的影响不具有显著调节作用。     

An extensive study on the evolution of context-aware personalized travel recommender systems

Ever since the beginning of civilization, travel for various causes exists as an essential part of human life so as travel recommendations, though the early form of recommendations were the accrued experiences shared by the community. Modern recommender systems evolved along with the growth of Information Technology and are contributing to all industry and service segments inclusive of travel and tourism. The journey started with generic recommender engines which gave way to personalized recommender systems and further advanced to contextualized personalization with advent of artificial intelligence. Current era is also witnessing a boom in social media usage and the social media big data is acting as a critical input for various analytics with no exception for recommender systems. This paper details about the study conducted on the evolution of travel recommender systems, their features and current set of limitations. We also discuss on the key algorithms being used for classification and recommendation processes and metrics that can be used to evaluate the performance of the algorithms and thereby the recommenders.     

Evaluating epistemic uncertainty under incomplete assessments

The thesis of this study is to propose an extended methodology for laboratory based Information Retrieval evaluation under incomplete relevance assessments. This new methodology aims to identify potential uncertainty during system comparison that may result from incompleteness. The adoption of this methodology is advantageous, because the detection of epistemic uncertainty – the amount of knowledge (or ignorance) we have about the estimate of a system’s performance – during the evaluation process can guide and direct researchers when evaluating new systems over existing and future test collections. Across a series of experiments we demonstrate how this methodology can lead towards a finer grained analysis of systems. In particular, we show through experimentation how the current practice in Information Retrieval evaluation of using a measurement depth larger than the pooling depth increases uncertainty during system comparison.     

Time and sequence awareness in similarity metrics for recommendation

Modeling the temporal context efficiently and effectively is essential to provide useful recommendations to users. In this work, we focus on improving neighborhood-based approaches where we integrate three different mechanisms to exploit temporal information. We first present an improved version of a similarity metric between users using a temporal decay function, then, we propose an adaptation of the Longest Common Subsequence algorithm to be used as a time-aware similarity metric, and we also redefine the neighborhood-based recommenders to be interpreted as ranking fusion techniques where the neighbor interaction sequence can be exploited by considering the last common interaction between the neighbor and the user.We demonstrate the effectiveness of these approaches by comparing them with other state-of-the-art recommender systems such as Matrix Factorization, Neural Networks, and Markov Chains under two realistic time-aware evaluation methodologies (per user and community-based). We use several evaluation metrics to measure both the quality of the recommendations – in terms of ranking relevance – and their temporal novelty or freshness. According to the obtained results, our proposals are highly competitive and obtain better results than the rest of the analyzed algorithms, producing improvements under the two evaluation dimensions tested consistently through three real-world datasets.     

A Pseudo-relevance feedback framework combining relevance matching and semantic matching for information retrieval

Pseudo-relevance feedback (PRF) is a well-known method for addressing the mismatch between query intention and query representation. Most current PRF methods consider relevance matching only from the perspective of terms used to sort feedback documents, thus possibly leading to a semantic gap between query representation and document representation. In this work, a PRF framework that combines relevance matching and semantic matching is proposed to improve the quality of feedback documents. Specifically, in the first round of retrieval, we propose a reranking mechanism in which the information of the exact terms and the semantic similarity between the query and document representations are calculated by bidirectional encoder representations from transformers (BERT); this mechanism reduces the text semantic gap by using the semantic information and improves the quality of feedback documents. Then, our proposed PRF framework is constructed to process the results of the first round of retrieval by using probability-based PRF methods and language-model-based PRF methods. Finally, we conduct extensive experiments on four Text Retrieval Conference (TREC) datasets. The results show that the proposed models outperform the robust baseline models in terms of the mean average precision (MAP) and precision P at position 10 (P@10), and the results also highlight that using the combined relevance matching and semantic matching method is more effective than using relevance matching or semantic matching alone in terms of improving the quality of feedback documents.     

What makes a review a reliable rating in recommender systems?

The way that users provide feedback on items regarding their satisfaction varies among systems: in some systems, only explicit ratings can be entered; in other systems textual reviews are accepted; and in some systems, both feedback types are accommodated. Recommender systems can readily exploit explicit ratings in the rating prediction and recommendation formulation process, however textual reviews -which in the context of many social networks are in abundance and significantly outnumber numeric ratings- need to be converted to numeric ratings. While numerous approaches exist that calculate a user's rating based on the respective textual review, all such approaches may introduce errors, in the sense that the process of rating calculation based on textual reviews involves an uncertainty level, due to the characteristics of the human language, and therefore the calculated ratings may not accurately reflect the actual ratings that the corresponding user would enter. In this work (1) we examine the features of textual reviews, which affect the reliability of the review-to-rating conversion procedure, (2) we compute a confidence level for each rating, which reflects the uncertainty level for each conversion process, (3) we exploit this metric both in the users’ similarity computation and in the prediction formulation phases in recommender systems, by presenting a novel rating prediction algorithm and (4) we validate the accuracy of the presented algorithm in terms of (i) rating prediction accuracy, using widely-used recommender systems datasets and (ii) recommendations generated for social network user satisfaction and precision, where textual reviews are abundant.     

Investigating and counteracting popularity bias in group recommendations

Popularity bias is an undesirable phenomenon associated with recommendation algorithms where popular items tend to be suggested over long-tail ones, even if the latter would be of reasonable interest for individuals. Such intrinsic tendencies of the recommenders may lead to producing ranked lists, in which items are not equally covered along the popularity tail. Although some recent studies aim to detect such biases of traditional algorithms and treat their effects on recommendations, the concept of popularity bias remains elusive for group recommender systems. Therefore, in this study, we focus on investigating popularity bias from the view of group recommender systems, which aggregate individual preferences to achieve recommendations for groups of users. We analyze various state-of-the-art aggregation techniques utilized in group recommender systems regarding their bias towards popular items. To counteract possible popularity issues in group recommendations, we adapt a traditional re-ranking approach that weighs items inversely proportional to their popularity within a group. Also, we propose a novel popularity bias mitigation procedure that re-ranks items by incorporating their popularity level and estimated group ratings in two distinct strategies. The first one aims to penalize popular items during the aggregation process highly and avoids bias better, while the second one puts more emphasis on group ratings than popularity and achieves a more balanced performance regarding conflicting goals of mitigating bias and boosting accuracy. Experiments performed on four real-world benchmark datasets demonstrate that both strategies are more efficient than the adapted approach, and empowering aggregation techniques with one of these strategies significantly decreases their bias towards popular items while maintaining reasonable ranking accuracy.     

Multilevel information system— Towards more flexible information retrieval systems

The Multilevel Information System (MLIS), extension of typical Information Retrieval System towards more complete data processing, is discussed. MLIS integrates functions typical for data base management systems and retrieval-oriented systems. Several levels of data accessing are provided, each level developed for a different class of users. End-user level is based on simple query language, trained user level on a relational model, and application programmer level on a Data Manipulation Language nested in high level programming language. The last two levels are discussed in detail.     

A new robust relevance model in the language model framework

In this paper, a new robust relevance model is proposed that can be applied to both pseudo and true relevance feedback in the language-modeling framework for document retrieval. There are at least three main differences between our new relevance model and other relevance models. The proposed model brings back the original query into the relevance model by treating it as a short, special document, in addition to a number of top-ranked documents returned from the first round retrieval for pseudo feedback, or a number of relevant documents for true relevance feedback. Second, instead of using a uniform prior as in the original relevance model proposed by Lavrenko and Croft, documents are assigned with different priors according to their lengths (in terms) and ranks in the first round retrieval. Third, the probability of a term in the relevance model is further adjusted by its probability in a background language model. In both pseudo and true relevance cases, we have compared the performance of our model to that of the two baselines: the original relevance model and a linear combination model. Our experimental results show that the proposed new model outperforms both of the two baselines in terms of mean average precision.     

Multidimensional relevance: Prioritized aggregation in a personalized Information Retrieval setting

A new model for aggregating multiple criteria evaluations for relevance assessment is proposed. An Information Retrieval context is considered, where relevance is modeled as a multidimensional property of documents. The usefulness and effectiveness of such a model are demonstrated by means of a case study on personalized Information Retrieval with multi-criteria relevance. The following criteria are considered to estimate document relevance: aboutness, coverage, appropriateness, and reliability.     

Query modification based on relevance back-propagation in an ad hoc environment

It is well-known that relevance feedback is a method significant in improving the effectiveness of information retrieval systems. Improving effectiveness is important since these information retrieval systems must gain access to large document collections distributed over different distant sites. As a consequence, efforts to retrieve relevant documents have become significantly greater. Relevance feedback can be viewed as an aid to the information retrieval task. In this paper, a relevance feedback strategy is presented. The strategy is based on back-propagation of the relevance of retrieved documents using an algorithm developed in a neural approach. This paper describes a neural information retrieval model and emphasizes the results obtained with the associated relevance back-propagation algorithm in three different environments: manual ad hoc, automatic ad hoc and mixed ad hoc strategy (automatic plus manual ad hoc).     

Probabilistic learning for selective dissemination of information

New methods and new systems are needed to filter or to selectively distribute the increasing volume of electronic information being produced nowadays. An effective information filtering system is one that provides the exact information that fulfills user's interests with the minimum effort by the user to describe it. Such a system will have to be adaptive to the user changing interest. In this paper we describe and evaluate a learning model for information filtering which is an adaptation of the generalized probabilistic model of Information Retrieval. The model is based on the concept of `uncertainty sampling', a technique that allows for relevance feedback both on relevant and nonrelevant documents. The proposed learning model is the core of a prototype information filtering system called ProFile.     

Improving the learning of Boolean queries by means of a multiobjective IQBE evolutionary algorithm

The Inductive Query By Example (IQBE) paradigm allows a system to automatically derive queries for a specific Information Retrieval System (IRS). Classic IRSs based on this paradigm [Smith, M., & Smith, M. (1997). The use of genetic programming to build Boolean queries for text retrieval through relevance feedback. Journal of Information Science, 23(6), 423–431] generate a single solution (Boolean query) in each run, that with the best fitness value, which is usually based on a weighted combination of the basic performance criteria, precision and recall.     

Dual disentanglement of user–item interaction for recommendation with causal embedding

To achieve personalized recommendations, the recommender system selects the items that users may like by learning the collected user–item interaction data. However, the acquisition and use of data usually form a feedback loop, which leads to recommender systems suffering from popularity bias. To solve this problem, we propose a novel dual disentanglement of user–item interaction for recommendation with causal embedding (DDCE). Different from the existing work, our innovation is we take into account double-end popularity bias from the user-side and the item-side. Firstly, we perform a causal analysis of the reasons for user–item interaction and obtain the causal embedding representation of each part according to the analysis results. Secondly, on the item-side, we consider the influence of item attributes on popularity to improve the reliability of the item popularity. Then, on the user-side, we consider the effect of the time series when obtaining users’ interest. We model the contrastive learning task to disentangle users’ long–short-term interests, which avoids the bias of long–short-term interests overlapping, and use the attention mechanism to realize the dynamic integration of users’ long–short-term interests. Finally, we realize the disentanglement of user–item interaction reasons by decoupling user interest and item popularity. We experiment on two real-world datasets (Douban Movie and KuaiRec) to verify the significance of DDCE, the average improvement of DDCE in three evaluation metrics (NDCG, HR, and Recall) compared to the state-of-the-art model are 5.1106% and 4.1277% (MF as the backbone), 3.8256% and 3.2790% (LightGCN as the backbone), respectively.     

Efficient federated item similarity model for privacy-preserving recommendation

Previous federated recommender systems are based on traditional matrix factorization, which can improve personalized service but are vulnerable to gradient inference attacks. Most of them adopt model averaging to fit the data heterogeneity of federated recommender systems, requiring more training costs. To address privacy and efficiency, we propose an efficient federated item similarity model for the heterogeneous recommendation, called FedIS, which can train a global item-based collaborative filtering model to eliminate user feature dependencies. Specifically, we extend the neural item similarity model to the federated model, where each client only locally optimizes the shared item feature matrix. We then propose a fast-convergent federated aggregation method inspired by meta-learning to address heterogeneous user updates and accelerate the convergence of global training. Furthermore, we propose a two-stage perturbation method to protect both local training and transmission while reducing communication costs. Finally, extensive experiments on four real-world datasets validate that FedIS can provide more competitive performance on federated recommendations. Our proposed method also shows significant training efficiency with less performance degradation.     

应用于信息检索的统计语言模型研究进展

统计语言模型作为一种自然语言处理的工具,已经被证明有能力处理大规模真实文本.而统计语言模型和IR相结合后所形成的SLM-IR模型的提出,是信息检索模型研究上的重大进展.本文介绍了统计语言模型在信息检索领域的基本模型及相关问题,重点分析了Lemur工具箱和标题语言模型的原理及模型,最后从整体上介绍了该领域的国际动态和研究进展情况.     

CLAIRE: A combinatorial visual analytics system for information retrieval evaluation

Information Retrieval (IR) develops complex systems, constituted of several components, which aim at returning and optimally ranking the most relevant documents in response to user queries. In this context, experimental evaluation plays a central role, since it allows for measuring IR systems effectiveness, increasing the understanding of their functioning, and better directing the efforts for improving them. Current evaluation methodologies are limited by two major factors: (i) IR systems are evaluated as “black boxes”, since it is not possible to decompose the contributions of the different components, e.g., stop lists, stemmers, and IR models; (ii) given that it is not possible to predict the effectiveness of an IR system, both academia and industry need to explore huge numbers of systems, originated by large combinatorial compositions of their components, to understand how they perform and how these components interact together.We propose a Combinatorial visuaL Analytics system for Information Retrieval Evaluation (CLAIRE) which allows for exploring and making sense of the performances of a large amount of IR systems, in order to quickly and intuitively grasp which system configurations are preferred, what are the contributions of the different components and how these components interact together.The CLAIRE system is then validated against use cases based on several test collections using a wide set of systems, generated by a combinatorial composition of several off-the-shelf components, representing the most common denominator almost always present in English IR systems. In particular, we validate the findings enabled by CLAIRE with respect to consolidated deep statistical analyses and we show that the CLAIRE system allows the generation of new insights, which were not detectable with traditional approaches.