Meta’s WHALE paper unifies Wukong-style interactions and HSTU sequence modeling
_reachsumit · x · 2026-07-21
Meta’s WHALE paper proposes a scalable unified recommendation model based on the Wukong-HSTU architecture.
- The model combines Wukong-style high-order feature interactions with HSTU-style sequence modeling in every layer.
- The fusion is done through attention-based mechanisms.
- The post frames WHALE as a unified architecture for recommendation rather than a product announcement.
More from Research
- Deep Learning Course Adds Mechanistic Interpretability Lecture This Winter — CSProfKGD · 2026-10-26
- INCLUDE-2.0 benchmark evaluates regional knowledge across 91 languages and 86 countries — ABosselut · 2026-10-26
- CoRL paper: flow-matching VLAs can run 1-step Euler inference with no retraining — kaixhin · 2026-10-26
- CoVer at EMNLP 2026: Training Code LLMs to Generate and Self-Verify Simultaneously — algo_diver · 2026-10-26
- Robotic dexterity should be judged on speed, force control and tool use, not copying humans — DJiafei · 2026-10-26
- Turing-complete malware can run in PyTorch's backward pass via a zero-valued loss — retr0jirachi · 2026-10-26