Alibaba's IntentGR strengthens query intent across four stages of personalized generative retrieval
_reachsumit · x · 2026-11-18
Alibaba researchers propose IntentGR, an intent-enhanced end-to-end framework for personalized generative retrieval, addressing how prior methods treat the query as a mere input feature, causing semantic IDs to drift toward high-frequency patterns.
- Strengthens query semantics at four complementary stages: Query-Aware Semantic Behavior Representation (QBR) adds query-level behavioral evidence; Query-Semantic Alignment (QSA), Long-Tail Query Enrichment (LQE) and Query-conditioned Attention Modulation (QAM) act at optimization, training and generation
- Shows notable gains on long-tail queries while jointly satisfying immediate intent and long-term preference
More from Research
- Voyager 1 Reaches One Light-Day From Earth, Signals Now Take 24 Hours — snowclipsed · 2026-11-18
- TCS Hiring Statement Urges Quality Over Quantity as AI Fuels Paper Arms Race — fortnow · 2026-11-18
- New paper: a sociocultural approach to linguistic creativity — abenitezburraco · 2026-11-18
- 27B Model Rewrites Its Own Weights Without Training or Gradients, Paper Claims — udmrzn · 2026-11-18
- From 'impossible' to a working Bluetooth beacon: open-source EDA chip design is real — IgorCarron · 2026-11-18
- AURA 2.0: a streaming video LLM that speaks first, calls tools, and grounds visually — Shuhan Li · 2026-11-18