Yandex's Long-History Ad Ranking Model
_reachsumit · x · 2026-07-17
Yandex has introduced a two-stage Transformer designed for real-time ad ranking:
- Offline Stage: Asynchronously encodes users' long interaction histories to leverage comprehensive long-term behavioral data.
- Online Stage: A lightweight runtime model processes the latest events to ensure real-time responsiveness.
The core focus of this architecture is to decouple "long-history modeling" from "low-latency inference," effectively balancing overall model performance with online execution speed.
More from Research
- Structural ensembles beat single predictions in TCR:pMHC generalization study — quaidmorris · 2026-07-22
- LLM leaderboards are now often measuring the harness too, Gary Marcus warns — GaryMarcus · 2026-07-22
- New paper defines self-state attacks, showing OS defenses leave four agent-memory cases indistinguishable — Justgototheeffinmoon · 2026-07-22
- Krea 2 users recommend a two-pass Clownshark sampler setup for sharper image details — listopalafoto · 2026-07-22
- Animation shows how an MLP’s first-layer weights change while learning MNIST — CatAstro_Piyush · 2026-07-22
- Project APE finds verifier reliability drops when papers contain multiple errors — soumitrashukla9 · 2026-07-22