A5022499183- Recommender Systems and Techniques
- Advanced Graph Neural Networks
- Consumer Market Behavior and Pricing
- Topic Modeling
- Data Mining Algorithms and Applications
- Machine Learning and Data Classification
- Multimodal Machine Learning Applications
- Caching and Content Delivery
- Data Management and Algorithms
- Advanced Bandit Algorithms Research
- Domain Adaptation and Few-Shot Learning
- Complex Network Analysis Techniques
Peking University
2023-2024
Session-based recommendation (SBR) aims to predict the user's next action based on short and dynamic sessions. Recently, there has been an increasing interest in utilizing various elaborately designed graph neural networks (GNNs) capture pair-wise relationships among items, seemingly suggesting design of more complicated models is panacea for improving empirical performance. However, these achieve relatively marginal improvements with exponential growth model complexity. In this paper, we...
With the rapid development of World Wide Web (WWW), heterogeneous graphs (HG) have explosive growth. Recently, graph neural network (HGNN) has shown great potential in learning on HG. Current studies HGNN mainly focus some HGs with strong homophily properties (nodes connected by meta-path tend to same labels), while few discussions are made those that less homophilous. there been many works homogeneous heterophily. However, due heterogeneity, it is non-trivial extend their approach deal In...
Collaborative filtering (CF) is an important research direction in recommender systems that aims to make recommendations given the information on user-item interactions. Graph CF has attracted more and attention recent years due its effectiveness leveraging high-order bipartite graph for better recommendations. Specifically, studies show success of neural networks (GNN) attributed low-pass effects. However, current researches lack a study how different signal components contributes...
Knowledge distillation is often used to transfer knowledge from a strong teacher model relatively weak student model. Traditional methods include response-based and feature-based methods. Response-based are widely but suffer lower upper limits of performance due their ignorance intermediate signals, while have constraints on vocabularies, tokenizers architectures. In this paper, we propose liberal method (LEAD). LEAD aligns the distribution between layers model, which effective, extendable,...
Auto-bidding plays a crucial role in facilitating online advertising by automatically providing bids for advertisers. Reinforcement learning (RL) has gained popularity auto-bidding. However, most current RL auto-bidding methods are modeled through the Markovian Decision Process (MDP), which assumes state transition. This assumption restricts ability to perform long horizon scenarios and makes model unstable when dealing with highly random environments. To tackle this issue, paper introduces...
Session-based recommendation (SBR) systems, traditionally reliant on complex graph neural networks (GNNs), often face challenges with marginal performance improvements despite increased model complexity. In this paper, we dissect the classical GNN-based SBR models and empirically find that sophisticated GNN propagations might be redundant, given readout module plays a significant role in models. Based observation, introduce Atten-Mixer+, an advanced iteration of our previously developed...