Tag
Kuaishou researchers propose UniGD, a unified generative-discriminative framework for industrial retrieval that integrates retrieval and relevance scoring into a single model, with techniques like CAGE and CAM to improve effectiveness and reduce latency. Online A/B tests show a 5.78% ad revenue increase and 33.1% inference latency reduction.
A 600M parameter reasoning model trained using SYNTH reportedly outperforms a 397B model and Sonnet 4.5 in an industrial application for the Paris subway, highlighting the effectiveness of small, specialized models.
ProfiLLM introduces an agentic LLM pipeline that generates utility-aligned user profiles from platform-scale behavioral logs for industrial ride-hailing dispatch, achieving significant improvements in outcome prediction and GMV in production at DiDi.