Research map / Large language models / Language, safety and interpretability
Amazon-M2: A Multilingual Multi-locale Shopping Session Dataset for Recommendation and Text Generation
Wei Jin, Haitao Mao, Zheng Li, Haoming Jiang, Chen Luo, Hongzhi Wen, Haoyu Han, Hanqing Lu, Zhengyang Wang, Ruirui Li, Zhen Li, Monica Xiao Cheng and 8 more
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 popularity heuristic instead of deep session models | session recommenders · Task 1 Overall | MRR@100: 0.1757-0.2159 -> 0.2426 | Improved |
| use CORE instead of popularity | session recommenders · Task 1 Overall | Recall@100: 0.4243 -> 0.5591 | Improved |
| pretrain on large locales and finetune on ES/FR/IT | SRGNN · Task 2 Overall | MRR@100: 0.2640 -> 0.2779 | Improved |
| pretrain and finetune | NARM · Task 2 Overall | Recall@100: 0.5920 -> 0.6127 | Improved |
| pretrain and finetune | CORE · Task 2 Overall | MRR@100: 0.2045 -> 0.1720 | Hurt |
| pretrain and finetune | STAMP · Task 2 Overall | Recall@100: 0.5217 -> 0.4803 | Hurt |
| initialize item embeddings with multilingual Sentence-BERT title embeddings (ablation) | SRGNN · Task 2 Overall | MRR@100: 0.2640 -> 0.2303 | Hurt |
| initialize item embeddings with Sentence-BERT title embeddings (ablation) | GRU4Rec++ · Task 2 Overall | Recall@100: 0.5856 -> 0.5454 | Hurt |
| increase session history length K from 1 to 3 (ablation) | mT5-small · Task 3 phase-1 test | BLEU: 0.2265 -> 0.2142 | No clear effect |
| increase model size from mT5-small to mT5-base (ablation) | mT5, K=1 · Task 3 phase-1 test | BLEU: 0.2265 -> 0.2251 | No clear effect |
| use Last Product Title heuristic instead of finetuned mT5 | title generation · Task 3 phase-1 test | BLEU: 0.2265 -> 0.2677 | Improved |
Closest papers by meaning
- Shopping MMLU: A Massive Multi-Task Online Shopping Benchmark for Large Language Models (NeurIPS 2024 · similarity 0.87)
- Language Representations Can be What Recommenders Need: Findings and Potentials (ICLR 2025 · similarity 0.84)
- LLM-ESR: Large Language Models Enhancement for Long-tailed Sequential Recommendation (NeurIPS 2024 · similarity 0.84)
- Ready for You When You Are Back: Content-Driven Session-Based Recommendation for Continuity of Experience (AAAI 2025 · similarity 0.84)
- Towards Unified Multi-Modal Personalization: Large Vision-Language Models for Generative Recommendation and Beyond (ICLR 2024 · similarity 0.84)
- Re2LLM: Reflective Reinforcement Large Language Model for Session-based Recommendation (AAAI 2025 · similarity 0.84)
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.
