Xiaohongshu's RedMall: generative e-commerce recommendations with hierarchical SIDs and RL
_reachsumit · x · 2026-11-10
Xiaohongshu researchers present RedMall, a generative recommender for large-scale e-commerce. It tackles three gaps in prior work: semantic IDs treated as flat sequences, next-token prediction not optimizing session-wise recommendation, and RL lacking explicit ranking optimization.
RedMall builds hierarchical SIDs via residual quantization over multimodal representations, adds a lightweight SID encoder for layer-wise dependencies, and scales with sparse MoE, dual gated attention, and multi-target prediction. An exposure-aware loss treats exposed-but-unclicked items as negatives, combining GRPO with an auxiliary LambdaRank objective for list-wise ranking.
More from Research
- Researchers debate optimizer equivariance: argument without orthogonality is more convincing — jasondeanlee · 2026-11-10
- Claude Fable 5.2 tops VIALS-AA science visual reasoning benchmark at 67% with heavy tool use — ArtificialAnlys · 2026-11-10
- WVU researcher presents 10 years of tractometry work on Alzheimer's white matter vulnerability — PTenigma · 2026-11-10
- New cryo-EM structure: pSTAT1 bound to importin alpha 5 (PDB 11GJ) — chaitjo · 2026-11-10
- ARLoop: fully-open 1.2B looped transformer beats Huginn-3.5B and Qwen3.5-2B on all 8 benchmarks — hhsun1 · 2026-11-10
- RAGCap-Bench: A Fine-Grained Benchmark for Intermediate Capabilities in Agentic RAG — kalyan_kpl · 2026-11-10