Google Research Introduces Retrieve-for-Train (R4T): An RL-Compiled Diffusion Retriever for 12× to 20× Faster Query Fan-Out
Google Research introduced R4T, an RL-trained fan-out pipeline distilled into a 53.9M-parameter diffusion retriever achieving 12x-20x faster query fan-out.
Google Research introduced Retrieve-for-Train (R4T), which trains a fan-out language model with GRPO plus soft PPO regularization, then distills query fan-out into a 53.9M-parameter diffusion transformer that generates all retrieval embeddings in a single non-autoregressive pass. A three-term reward (groundedness 0.6, diversity 0.2 via Vendi Score, alignment 0.2) prevents paraphrastic collapse and reward hacking during training. On the Polyvore dataset, Gemma3-4B R4T-FOLM averaged 49.1 versus 40.9 for Best-of-N, and the diffusion retriever cut fan-out latency from 1.46s to 0.07s at batch size 8, a consistent 12x-20x speedup over autoregressive methods.
- R4T runs RL once offline, then deploys a small diffusion retriever for set-level query fan-out.
- Groundedness, diversity, and alignment rewards jointly block reward hacking and paraphrastic collapse.
- Polyvore OAR average improved from 40.9 (Best-of-N) to 49.1 with Gemma3-4B; music dataset rose from 49.2 to 58.1.
- Supervision is synthesized from 128 fan-outs per query, requiring no human labels.
- Vendors
- Organizations
- Google Research
- AI models
- R4TGemma3-4BQwen3-4BGemini-2.5-Flash
From the original article
4 min read at marktechpost.com
Google Research has introduced Retrieve-for-Train (R4T), a framework for search that returns coherent, diverse result sets. It trains a fan-out language model with RL once, using groundedness, diversity, and alignment rewards. That model then synthesizes training data for a 53.
By Asif Razzaq · marktechpost.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.