GenRec combines page-wise NTP, token compression, and GRPO-SR reinforcement learning to scale generative retrieval, delivering 9.5% click and 8.7% transaction gains in production A/B tests on the JD App.
arXiv preprint arXiv:2510.26104 , year=
12 Pith papers cite this work. Polarity classification is still indexing.
12
Pith papers citing it
years
2026 12representative citing papers
TokenFormer unifies multi-field and sequential recommendation modeling via bottom-full-top-sliding attention and non-linear interaction representations to avoid sequential collapse and deliver state-of-the-art performance.
IAT compresses each historical interaction instance into a unified embedding token via temporal-order or user-order schemes, allowing standard sequence models to learn long-range preferences with better performance and transferability.
MTServe achieves up to 3.1x speedup for generative recommendation model serving by using hierarchical caches with host RAM and system optimizations while keeping cache hit ratios above 98.5%.
UniSID jointly optimizes embeddings and Semantic IDs end-to-end with multi-granularity contrastive learning and summary-based reconstruction, outperforming RQ-based methods by up to 4.62% in Hit Rate for ad recommendation.
SSRLive combines generative and discriminative modules with dynamic semantic IDs to improve live streaming recommendations, reporting gains of +3.38% watch time, +0.72% GMV, +3.12% follower growth, and +2.92% interaction volume in online A/B tests.
GrowthGR combines ItemLTV counterfactual prediction with MultiGR generative retrieval and MoPO optimization to deliver 5.3% new item GMV lift and 0.3% overall GMV gain on Taobao production.
SIF encodes entire historical raw samples as tokens via hierarchical group-adaptive quantization and token/sample-level mixing to overcome partial encoding and feature heterogeneity limits in scaled recommender models.
A re-ranking system for rich-media search that plans query intents from sessions, adds visual signals from VLMs, and uses an LLM to score results on multiple facets before multi-task RL adaptation, with reported gains in engagement after industrial deployment.
CMSL uses a learnable module to disentangle user history into multiple pure sequences modeled with linear attention to improve recommendation performance over single-sequence approaches.
UniScale couples entire-space data construction with a hierarchical fusion transformer to improve scaling behavior and deliver 1.70% purchase and 2.04% GMV lifts in large-scale e-commerce search A/B tests.