本文针对序列推荐中文本特征与 ID 特征直接融合效果不佳的问题，提出 TedRec，通过傅里叶变换在频域进行乘法式序列级融合，并利用 MoE 注入位置信息以增强文本表示。
摘要：
随着推荐系统的快速发展，可用于提升推荐性能的辅助信息日益增多。特别地，我们关注物品相关文本数据（例如商品标题）的利用，并研究在序列推荐中如何将文本特征与 ID 特征有效融合。然而，这两种物品特征具有明显不同的数据特性，导致直接融合方法（例如将文本嵌入与 ID 嵌入相加作为物品表示）效果不佳。为解决该问题，我们提出了一种新颖的面向序列推荐的 Text-ID 语义融合方法，即 TedRec。该方法的核心思想是通过更好地整合全局上下文来实施序列级语义融合。关键策略在于：我们利用 Fourier Transform（傅里叶变换）将文本嵌入和 ID 嵌入从时域转换到频域。在频域中，原始序列的全局序列特性会被内在地聚合到变换后的表示中，因此我们可以使用简单的乘法运算有效融合这两类物品特征。我们的融合方法可被证明具有与上下文卷积相同的效果，从而实现序列级语义融合。为进一步提升融合性能，我们提出通过 mixture-of-experts（MoE，专家混合）调制方法自适应地注入位置信息，以增强来自文本编码器的文本嵌入的可判别性。我们的实现可在以下仓库获取：https://github.com/RUCAIBox/TedRec。
英文摘要：
With the rapid development of recommender systems, there is increasing side information that can be employed to improve the recommendation performance. Specially, we focus on the utilization of the associated \emph{textual data} of items (eg product title) and study how text features can be effectively fused with ID features in sequential recommendation. However, there exists distinct data characteristics for the two kinds of item features, making a direct fusion method (eg adding text and ID embeddings as item representation) become less effective. To address this issue, we propose a novel {\ul \emph{Te}}xt-I{\ul \emph{D}} semantic fusion approach for sequential {\ul \emph{Rec}}ommendation, namely \textbf{\our}. The core idea of our approach is to conduct a sequence-level semantic fusion approach by better integrating global contexts. The key strategy lies in that we transform the text embeddings and ID embeddings by Fourier Transform from \emph{time domain} to \emph{frequency domain}. In the frequency domain, the global sequential characteristics of the original sequences are inherently aggregated into the transformed representations, so that we can employ simple multiplicative operations to effectively fuse the two kinds of item features. Our fusion approach can be proved to have the same effects of contextual convolution, so as to achieving sequence-level semantic fusion. In order to further improve the fusion performance, we propose to enhance the discriminability of the text embeddings from the text encoder, by adaptively injecting positional information via a mixture-of-experts~(MoE) modulation method. Our implementation is available at this repository: \textcolor{magenta}{\url{https://github.com/RUCAIBox/TedRec}}.
We describe the Universal Recommender, a recommender system for semantic datasets that generalizes domain-specific recommenders such as content-based, collaborative, social, bibliographic, lexicographic, hybrid and other recommenders. In contrast to existing recommender systems, the Universal Recommender applies to any dataset that allows a semantic representation. We describe the scalable three-stage architecture of the Universal Recommender and its application to Internet Protocol Television (IPTV). To achieve good recommendation accuracy, several novel machine learning and optimization prob…
Recommending Chemical Compounds of interest to a particular researcher is a poorly explored field. The few existent datasets with information about the preferences of the researchers use implicit feedback. The lack of Recommender Systems in this particular field presents a challenge for the development of new recommendations models. In this work, we propose a Hybrid recommender model for recommending Chemical Compounds. The model integrates collaborative-filtering algorithms for implicit feedback (Alternating Least Squares (ALS) and Bayesian Personalized Ranking(BPR)) and semantic similarity b…
We assume that recommender systems are more successful, when they are based on a thorough understanding of how people process information. In the current paper we test this assumption in the context of social tagging systems. Cognitive research on how people assign tags has shown that they draw on two interconnected levels of knowledge in their memory: on a conceptual level of semantic fields or topics, and on a lexical level that turns patterns on the semantic level into words. Another strand of tagging research reveals a strong impact of time dependent forgetting on users' tag choices, such …
Recommender systems are needed to find food items of ones interest. We review recommender systems and recommendation methods. We propose a food personalization framework based on adaptive hypermedia. We extend Hermes framework with food recommendation functionality. We combine TF-IDF term extraction method with cosine similarity measure. Healthy heuristics and standard food database are incorporated into the knowledgebase. Based on the performed evaluation, we conclude that semantic recommender systems in general outperform traditional recommenders systems with respect to accuracy, precision, …
Recommender systems usually operate on similarities between recommended items or users. Tag based recommender systems utilize similarities on tags. The tags are however mostly free user entered phrases. Therefore, similarities computed without their semantic groundings might lead to less relevant recommendations. In this paper, we study a semantic grounding used for tag similarity calculus. We show a comprehensive analysis of semantic grounding given by 20 ontologies from different domains. The study besides other things reveals that currently available OWL ontologies are very narrow and the p…
Peking University（北京大学）；Renmin University of China（中国人民大学）
Conversational recommender systems (CRS) aim to recommend high-quality items to users through interactive conversations. Although several efforts have been made for CRS, two major issues still remain to be solved. First, the conversation data itself lacks of sufficient contextual information for accurately understanding users' preference. Second, there is a semantic gap between natural language expression and item-level user preference. To address these issues, we incorporate both word-oriented and entity-oriented knowledge graphs (KG) to enhance the data representations in CRSs, and adopt Mut…
Online forums enable users to discuss together around various topics. One of the serious problems of these environments is high volume of discussions and thus information overload problem. Unfortunately without considering the users interests, traditional Information Retrieval (IR) techniques are not able to solve the problem. Therefore, employment of a Recommender System (RS) that could suggest favorite's topics of users according to their tastes could increases the dynamism of forum and prevent the users from duplicate posts. In addition, consideration of semantics can be useful for increasi…
Due the success of emerging Web 2.0, and different social network Web sites such as Amazon and movie lens, recommender systems are creating unprecedented opportunities to help people browsing the web when looking for relevant information, and making choices. Generally, these recommender systems are classified in three categories: content based, collaborative filtering, and hybrid based recommendation systems. Usually, these systems employ standard recommendation methods such as artificial neural networks, nearest neighbor, or Bayesian networks. However, these approaches are limited compared to…
Synerise S.A.（Synerise股份公司）；Warsaw University of Technology（华沙理工大学）
Recommendation systems have lately been popularized globally, with primary use cases in online interaction systems, with significant focus on e-commerce platforms. We have developed a machine learning-based recommendation platform, which can be easily applied to almost any items and/or actions domain. Contrary to existing recommendation systems, our platform supports multiple types of interaction data with multiple modalities of metadata natively. This is achieved through multi-modal fusion of various data representations. We deployed the platform into multiple e-commerce stores of different k…
This paper explores the use of hyperbolic geometry and deep learning techniques for recommendation. We present Hyperbolic Neural Collaborative Recommender (HNCR), a deep hyperbolic representation learning method that exploits mutual semantic relations among users/items for collaborative filtering (CF) tasks. HNCR contains two major phases: neighbor construction and recommendation framework. The first phase introduces a neighbor construction strategy to construct a semantic neighbor set for each user and item according to the user-item historical interaction. In the second phase, we develop a d…
Collaborative filtering (CF) has been one of the most important and popular recommendation methods, which aims at predicting users' preferences (ratings) based on their past behaviors. Recently, various types of side information beyond the explicit ratings users give to items, such as social connections among users and metadata of items, have been introduced into CF and shown to be useful for improving recommendation performance. However, previous works process different types of information separately, thus failing to capture the correlations that might exist across them. To address this prob…
Capturing users' precise preferences is a fundamental problem in large-scale recommender system. Currently, item-based Collaborative Filtering (CF) methods are common matching approaches in industry. However, they are not effective to model dynamic and evolving preferences of users. In this paper, we propose a new sequential deep matching (SDM) model to capture users' dynamic preferences by combining short-term sessions and long-term behaviors. Compared with existing sequence-aware recommendation methods, we tackle the following two inherent problems in real-world applications: (1) there could…
Semantic-Enhanced Relational Metric Learning for Recommender Systems
Institute of Information Engineering, Chinese Academy of Sciences（中国科学院信息工程研究所）；School of Cyber Security, University of Chinese Academy of Sciences（中国科学院大学网络空间安全学院）；School of Computer Science and Engineering, The University of New South Wales（新南威尔士大学计算机科学与工程学院）
Recommender System (RS) is an important online application that affects billions of users every day. The mainstream RS ranking framework is composed of two parts: a Multi-Task Learning model (MTL) that predicts various user feedback, i.e., clicks, likes, sharings, and a Multi-Task Fusion model (MTF) that combines the multi-task outputs into one final ranking score with respect to user satisfaction. There has not been much research on the fusion model while it has great impact on the final recommendation as the last crucial process of the ranking. To optimize long-term user satisfaction rather …
Knowledge graph (KG) plays an increasingly important role in recommender systems. Recently, graph neural networks (GNNs) based model has gradually become the theme of knowledge-aware recommendation (KGR). However, there is a natural deficiency for GNN-based KGR models, that is, the sparse supervised signal problem, which may make their actual performance drop to some extent. Inspired by the recent success of contrastive learning in mining supervised signals from data itself, in this paper, we focus on exploring the contrastive learning in KG-aware recommendation and propose a novel multi-level…
Deep learning-based recommender systems may lead to over-fitting when lacking training interaction data. This over-fitting significantly degrades recommendation performances. To address this data sparsity problem, cross-domain recommender systems (CDRSs) exploit the data from an auxiliary source domain to facilitate the recommendation on the sparse target domain. Most existing CDRSs rely on overlapping users or items to connect domains and transfer knowledge. However, matching users is an arduous task and may involve privacy issues when data comes from different companies, resulting in a limit…
Recommender systems leverage both content and user interactions to generate recommendations that fit users' preferences. The recent surge of interest in deep learning presents new opportunities for exploiting these two sources of information. To recommend items we propose to first learn a user-independent high-dimensional semantic space in which items are positioned according to their substitutability, and then learn a user-specific transformation function to transform this space into a ranking according to the user's past preferences. An advantage of the proposed architecture is that it can b…
