A5008399761- Recommender Systems and Techniques
- Advanced Graph Neural Networks
- Topic Modeling
- Advanced Bandit Algorithms Research
- Data Management and Algorithms
- Multi-Criteria Decision Making
- Graph Theory and Algorithms
- Manufacturing Process and Optimization
- Information Retrieval and Search Behavior
- Product Development and Customization
- Machine Learning in Healthcare
Renmin University of China
2018-2023
Alibaba Group (China)
2023
In recent years, there are a large number of recommendation algorithms proposed in the literature, from traditional collaborative filtering to deep learning algorithms. However, concerns about how standardize open source implementation continually increase research community. light this challenge, we propose unified, comprehensive and efficient recommender system library called RecBole (pronounced as [rEk'[email protected]]), which provides unified framework develop reproduce for purpose....
In order to develop effective sequential recommenders, a series of sequence representation learning (SRL) methods are proposed model historical user behaviors. Most existing SRL rely on explicit item IDs for developing the models better capture preference. Though some extent, these difficult be transferred new recommendation scenarios, due limitation by explicitly modeling IDs. To tackle this issue, we present novel universal approach, named UniSRec. The approach utilizes associated...
In order to support the study of recent advances in recommender systems, this paper presents an extended recommendation library consisting eight packages for up-to-date topics and architectures. First all, from a data perspective, we consider three important related issues (ie sparsity, bias distribution shift ), develop five accordingly, including meta-learning, augmentation, debiasing, fairness cross-domain recommendation. Furthermore, model two benchmarking Transformer-based graph neural...
Limited by the statistical-based machine learning framework, a spurious correlation is likely to appear in existing knowledge-aware recommendation methods. It refers knowledge fact that appears causal user behaviors (inferred recommender) but not fact. For tackling this issue, we present novel approach discovering and alleviating potential correlations from counterfactual perspective. To be specific, our consists of two generators recommender. The are designed generate interactions via...
In recent years, there are a large number of recommendation algorithms proposed in the literature, from traditional collaborative filtering to deep learning algorithms. However, concerns about how standardize open source implementation continually increase research community. light this challenge, we propose unified, comprehensive and efficient recommender system library called RecBole, which provides unified framework develop reproduce for purpose. library, implement 73 models on 28...
As it becomes prevalent that user information exists in multiple platforms or services, cross-domain recommendation has been an important task industry. Although is well known users tend to show different preferences domains, existing studies seldom model how domain biases affect preferences. Focused on this issue, we develop a casual-based approach mitigating the when transferring cross domains. To be specific, paper presents novel debiasing learning based framework with causal embedding....
In recommender systems, it is essential to understand the underlying factors that affect user-item interaction. Recently, several studies have utilized disentangled representation learning discover such hidden from interaction data, which shows promising results. However, without any external guidance signal, learned representations lack clear meanings, and are easy suffer data sparsity issue. light of these challenges, we study how leverage knowledge graph (KG) guide in systems. The purpose...
Multi-scenario ad ranking aims at leveraging the data from multiple domains or channels for training a unified model improving performance each individual scenario. Although research on this task has made important progress, it still lacks consideration of cross-scenario relations, thus leading to limitation in learning capability and difficulty interrelation modeling.
In order to develop effective sequential recommenders, a series of sequence representation learning (SRL) methods are proposed model historical user behaviors. Most existing SRL rely on explicit item IDs for developing the models better capture preference. Though some extent, these difficult be transferred new recommendation scenarios, due limitation by explicitly modeling IDs. To tackle this issue, we present novel universal approach, named UniSRec. The approach utilizes associated...
Multi-scenario ad ranking aims at leveraging the data from multiple domains or channels for training a unified model to improve performance each individual scenario. Although research on this task has made important progress, it still lacks consideration of cross-scenario relations, thus leading limitation in learning capability and difficulty interrelation modeling. In paper, we propose Hybrid Contrastive Constrained approach (HC^2) multi-scenario ranking. To enhance modeling interrelation,...
Recently, sequential recommendation has emerged as a widely studied topic. Existing researches mainly design effective neural architectures to model user behavior sequences based on item IDs. However, this kind of approach highly relies user-item interaction data and neglects the attribute- or characteristic-level correlations among similar items preferred by user. In light these issues, we propose IDA-SR, which stands for ID-Agnostic User Behavior Pre-training Sequential Recommendation....
In order to support the study of recent advances in recommender systems, this paper presents an extended recommendation library consisting eight packages for up-to-date topics and architectures. First all, from a data perspective, we consider three important related issues (i.e., sparsity, bias distribution shift), develop five accordingly: meta-learning, augmentation, debiasing, fairness cross-domain recommendation. Furthermore, model two benchmarking Transformer-based graph neural network...