Alibaba's Decoupled Temporal Encoding for Generative Recommendation Now Live in Taobao Ads
_reachsumit · x · 2026-08-18
Alibaba proposes the Decoupled Temporal Encoding (DTE) framework for Transformer-based generative recommendation models. The framework separates macro-level temporal dynamics (such as recency, mealtime peaks, and weekday/weekend patterns) from micro-level local sequential information, modeling them respectively through a personalized macro temporal module and a time-gated micro-sequence module. DTE is parameter-efficient and has been deployed in Taobao's advertising ranking system, effectively handling multi-level temporal patterns in food delivery and instant retail scenarios.
More from Research
- ImageNet should retire: DiT method rankings negatively correlate between ImageNet and text-to-image — YouJiacheng · 2026-10-16
- GAE paper learns a geometry-native latent space for 3D-consistent world generation — yshan2u · 2026-10-16
- Agents turn conservative as context grows: edits shrink from 150 to 30 lines — dbreunig · 2026-10-16
- LMR-BENCH at EMNLP 2025 tests whether LLM agents can reproduce real LLM research — ZhiyuChen4 · 2026-10-16
- Ex-OpenAI researcher claims most neurons are legible now; critics call it hype — anshulkundaje · 2026-10-16
- Physicist pushes back on ML community's acceptance that models may be beyond understanding — tszzl · 2026-10-16