2026-11-09 ~ 11-15 · Generative
Generative · Personalization · AI Search · Infra · Embodied AI
- 1. Generative Search and Recommendation at Shopify · 📖 Deep dive
Search and recommendation at Shopify have to retrieve relevant products from a changing catalog of billions of items under tight latency limits. This report casts both tasks as generating semantic IDs: short sequences…
Appeared in: 11-10
- 2. RedMall: Generative Recommendation for Large-Scale E-commerce · 📖 Deep dive
Generative retrieval outputs a Semantic ID (SID), but several products can share one SID, and which of them are finally retrieved directly affects online gains. RedMall is a generative model for e-commerce…
Appeared in: 11-09
- 3. TS-GEAR: Taobao Shangou GEnerative Ad Retrieval · 📖 Deep dive
In food delivery search, a merchant that cannot reach the user's current location is not merely ranked low, it must not appear at all. Common SIDs are built only from content and collaborative signals and at best treat…
Appeared in: 11-11
- 4. Not All Branches Are Equal: Adaptive Semantic ID Construction for Generative Recommendation · 📖 Deep dive
Generative recommendation numbers products with SIDs, and the dominant approach is residual quantization: each level has one global code table and looks up whatever the previous level left unexplained…
Appeared in: 11-11
- 5. Rethinking Latent Reasoning for Generative Recommendation: Beyond Teacher-Generated CoT and Recurrent Hidden-State Feedback · 📖 Deep dive
Latent reasoning for generative recommendation usually needs a teacher model to write chain-of-thought traces, which is costly and hard to stabilize. This paper skips teacher generation…
Appeared in: 11-10
- 6. IntentLens: Grounding Underspecified Multimodal Queries for Recommendation via Tool-Augmented Reasoning · 📖 Deep dive
On devices such as smart glasses, users often give only a picture and one vague sentence, leaving style, budget and venue requirements unsaid. Existing recommenders either rely on past behavior or assume the request is…
Appeared in: 11-11
- 7. RecGroupMemory: Feedback-Aware Cohort Memory for E-Commerce Recommendation · 📖 Deep dive
When a large model plans a recommendation workflow, it can be given memory of how a group of users has behaved. The difficulty is that new transaction feedback arrives the next day and the advice should change…
Appeared in: 11-09
- 8. LEGIT: Live Eligibility Gating over Identifier Tries for Constrained Decoding of Heterogeneous Pages
Uber Eats generative search writes a whole page of stores and menu items in one decode, and every token must keep each store real, deliverable and unrepeated, with each item belonging to the store above it…
Appeared in: 11-11
- 9. RoutingTrie: Metadata Routing for Constrained Generative Retrieval · 📖 Deep dive
Generative retrieval often has to honor a condition, such as returning only products from one shop, while a trie only guarantees that a generated SID exists and says nothing about shops or countries…
Appeared in: 11-11
- 10. CascadeRec: Deployed Adaptive LLM Compute Allocation for Multi-Tenant Content Recommendation · 📖 Deep dive
Large language models are attractive rerankers, but running an expensive model on every request is costly and often hurts users with short histories. This production system…
Appeared in: 11-10