Research map / Large language models / AI and society
DaisyRec 2.0: Benchmarking Recommendation for Rigorous Evaluation
Zhu Sun, Hui Fang, Jie Yang, Xinghua Qu, Hongyang Liu, Di Yu, Yew-Soon Ong, Jie Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use time-aware instead of random split-by-ratio | ten baselines · six datasets, 10-filter | NDCG@10 | Hurt |
| use leave-one-out instead of split-by-ratio | ten baselines · six datasets, 10-filter | accuracy | Hurt |
| tune hyper-parameters on test set instead of validation set (ablation) | BPRMF, BPRFM, NeuMF, NFM etc. · four datasets, 10-filter | NDCG@10 | Improved |
| use BPR loss instead of CE, Hinge, Top1 (ablation) | MF, FM, NeuMF, NFM, NGCF · six datasets, 10-filter | NDCG@10 | Improved |
| use CE loss instead of BPR (ablation) | NeuMF · six datasets, 10-filter | NDCG@10 | Improved |
| use uniform instead of popularity-based negative sampler (ablation) | BPRMF, BPRFM, NeuMF, NFM, NGCF · six datasets, 10-filter | NDCG@10 | No clear effect |
| mix uniform with popularity sampler (U+HP, U+LP) (ablation) | five baselines · six datasets, 10-filter | NDCG@10 | Hurt |
| use normal instead of uniform initializer (ablation) | BPRMF, BPRFM · six datasets, 10-filter | NDCG@10 | Improved |
| use SGD instead of Adam (ablation) | BPRMF, BPRFM · six datasets, 10-filter | NDCG@10 | Improved |
| use Adam instead of SGD (ablation) | NeuMF, NFM, NGCF, Multi-VAE · six datasets, 10-filter | NDCG@10 | Improved |
| add overfitting prevention (L2, dropout, early stop) (ablation) | six baselines · six datasets, 10-filter | NDCG@10 | Improved |
| use 5- or 10-filter instead of origin dataset | ten baselines · Book-X, Yelp, AMZe | NDCG@10 | Improved |
Closest papers by meaning
- ORBIT - Open Recommendation Benchmark for Reproducible Research with Hidden Tests (NeurIPS 2025 · similarity 0.89)
- RecFlow: An Industrial Full Flow Recommendation Dataset (ICLR 2025 · similarity 0.88)
- AgentRecBench: Benchmarking LLM Agent-based Personalized Recommender Systems (NeurIPS 2025 · similarity 0.86)
- Addressing Hidden Confounding with Heterogeneous Observational Datasets for Recommendation (NeurIPS 2024 · similarity 0.86)
- CURE4Rec: A Benchmark for Recommendation Unlearning with Deeper Influence (NeurIPS 2024 · similarity 0.86)
- Would You Like Your Data to Be Trained? A User Controllable Recommendation Framework (AAAI 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
