JevRec: variable-length semantic IDs boost LLM generative recommendation on three benchmarks
_reachsumit · x · 2026-10-29
The arXiv paper JevRec improves LLM-based generative recommendation with linguistic encoding and Jev-like decision modeling.
- Problem: most GR methods assign uniform-length semantic IDs (SIDs) despite hugely unequal item frequencies, wasting fine-grained codes on popular items while shortening codes compresses long-tail representations.
- Method: inspired by compact forms of frequent expressions in language, popular items get shorter SIDs and long-tail items richer descriptions; a dedicated head predicts sequence-level length distribution while a separate head generates content conditioned on length, resolving conflicting optimization objectives.
- Outperforms all baselines across three public datasets.
More from Research
- Granary-v2 open speech model expands to 48 languages with open data and pipeline — rdesh26 · 2026-10-29
- Tinker launches Dimuon, a Muon-based optimizer that beats AdamW on LoRA fine-tuning — xennygrimmato_ · 2026-10-29
- Six CUDA kernel optimizations squeeze another 1.78× speedup out of 3DGS training — gkopanas · 2026-10-29
- Cross-lingual LLM effects paper wins Outstanding Paper award at EMNLP 2026 — anas_ant · 2026-10-29
- Lie detector compiled into Qwen3.5 weights flags 91% of lies with zero false positives — teortaxesTex · 2026-10-29
- Practitioner: No serious lab actually trains on the test set today — nrehiew_ · 2026-10-29