Qijiong LiuHangzhou Dianzi University

Regenerating recommendation.

I build and evaluate recommender systems that use large models, with a focus on semantic tokenization, generative recommendation, and efficient inference.

Findings of EMNLP · 2026 Knowledge Bias in LLM-based Sequential Recommendation: Source, Persistence, and Mitigation

Research directions

Generative Recommendation
From ranking scores to generative systems.
Evaluation & Benchmarks
Testing whether models can truly recommend.
Efficient Modeling
Sharper representations, less computation.

Publications

All publications
  • Findings of EMNLP · 2026

    Knowledge Bias in LLM-based Sequential Recommendation: Source, Persistence, and Mitigation

  • WSDM · 2026

    Accelerating Generative Recommendation via Simple Categorical User Sequence Compression

  • SIGIR · 2026

    Full Retraining, Incremental Fine-tuning, and Hybrid Serving: Model Updating and Serving for Industrial Generative Recommender Systems

  • EReL@MIR · ACM Multimedia · 2026

    The 2nd EReL@MIR Workshop on Efficient Representation Learning for Multimodal Information Retrieval

  • PAKDD · 2026

    Learning Multi-aspect Item Palette: A Semantic Tokenization Framework for Generative Recommendation

Projects

Legommenders

About Qijiong Liu

Lixing Distinguished Professor · Hangzhou Dianzi University

Read resume