ByteDance's HELIX Unifies Sequence Retrieval and Feature Interaction, Deployed in TikTok
_reachsumit · x · 2026-09-30
ByteDance published HELIX, a new ranker architecture for large-scale recommendation, already deployed in TikTok's e-commerce system.
- Key finding: Scaling feature interaction or long sequence modeling in isolation hits limited ceilings with suboptimal scaling-law slopes; joint scaling of both axes is needed for a better slope.
- Design: HELIX interleaves sequence retrieval and feature interaction, enforcing one-way information flow from reusable sequence states to candidate-conditioned mix-tokens — preserving cross-depth communication while keeping user-side compute amortizable and enabling asymmetric scaling.
- Results: Consistent offline improvements in CTR AUC, CVR AUC, and other ranking metrics in TikTok's e-commerce recommendation.
More from Infra
- 8% of Asia-to-US air freight is now data center parts — 30 full freighters a day — yacineMTB · 2026-09-30
- mradermacher quants get Gemma 26B to 75 tok/s on 2x RTX 4060 8GB — Spiritual_Impress_30 · 2026-09-30
- Hugging Face ships tokenizers v1, often tens of times faster than v0.23 — ariG23498 · 2026-09-30
- Altman says OpenAI's custom chip program comes online in H1 2027, betting on inference advantage — johncoogan · 2026-09-30
- Rumor: DeepSeek's rumored single-GPU model may have been trained on Ascend — teortaxesTex · 2026-09-30
- AAOI burnt capex on US vertical integration, CW laser yield still poor — jwt0625 · 2026-09-30