Revolutionizing ASR in Regulated Domains with Reinforcement Learning
New research suggests that reinforcement learning, specifically GRPO, outperforms traditional fine-tuning for ASR in regulated sectors. This approach significantly lowers WER using synthetic speech.
Read the full articleYou might also wanna read
Study Finds Single Transformer Layer Can Match Full-Parameter RL Training in LLMs
Reinforcement learning (RL) has become a central component of post-training large language models (LLMs), yet little is understood about how
arxiv.org·8d agoStudy Finds Single Transformer Layer Can Match Full-Parameter RL Training in LLMs
Reinforcement learning (RL) has become a central component of post-training large language models (LLMs), yet little is understood about how
arxiv.org·8d ago
ConSPO: A Contrastive Approach to Improving Reinforcement Learning with Verifiable Rewards for LLMs
Group Relative Policy Optimization (GRPO) is one of the most widely adopted RLVR algorithms for post-training large language models on reaso
arxiv.org·1mo ago
10 Open-Source Reinforcement Learning Frameworks for Training AI Agents
A practical list of agent RL training frameworks for GRPO, multi-step agents, tool use, long-horizon tasks, rollouts, and multi-agent workfl
turingpost.com·17d ago
Feedback Distillation: A New Training Method for Improving LLM Reasoning in Theorem Proving
Post-training for reasoning models typically combines supervised fine-tuning with reinforcement learning from verifiable rewards, most commo
arxiv.org·1mo ago
SFT Overtraining Causes Rank Inversion in RLVR via Entropy Collapse
The standard heuristic of selecting the SFT checkpoint with the highest pass@1 for GRPO can fail when SFT compresses the rollout distributio
arxiv.org·17d agoIntroduction to Self-Adapting Language Models (SEAL)
Large language models (LLMs) are powerful but static; they lack mechanisms to adapt their weights in response to new tasks, knowledge, or ex
arxiv.org·1y ago
Comments
Sign in to join the conversation.
No comments yet. Be the first.