HASTE Adds Hierarchical and Attribute-aware Losses to Speed Up Generative Recommendation Training
_reachsumit · x · 2026-10-14
The HASTE paper (arXiv:2610.17368) improves generative recommendation models that autoregressively generate Semantic IDs, replacing plain causal cross-entropy:
- Hierarchical Loss supervises every level of the item hierarchy, exploiting the multi-level structure of SIDs and substantially accelerating convergence.
- Attribute-Specific Losses add attribute prediction heads that leverage an LM's inherent semantic knowledge, preserving fine-grained item attributes and improving representations for data-sparse tail items.
The authors report simultaneous gains in training efficiency and long-tail recommendation performance.
More from Research
- FrontierML benchmark tests whether agents can spend money to scale model performance — JayAlammar · 2026-10-14
- AIPOCH's Open-Science ships portable .science packages that share full research context, not just outputs — ahuja_priyank · 2026-10-14
- Video-Arena: Game-Based Video Benchmark Shows Native Models Score Just 13.2 F1 — Kuan Zhang · 2026-10-14
- LSCBench: Vibe Coding Agents Write Functional Code, But Only 25% Is Also Secure — Qwen · 2026-10-14
- Omni-LLMs Know More Than They Say: NAVER's OmniCatalyst Adds 7.29% via Self-Generated Text — naver-ai · 2026-10-14
- Repeated-Data Pretraining Study: Smaller Batches Act as Regularizers and Beat More GPUs — Sewon Min · 2026-10-14