Conduit: An Experience Data Plane for Distributed Reinforcement Learning
cs.DC, cs.AI
Submitted: 2026-09-21
Updated: 2026-09-21
Comments: 16 pages, 17 figures
Code: https://github.com/google-deepmind/reverb
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- AReaL: A Large-Scale Asynchronous Reinforcement Learning System for Language Reasoning
- HetRL: Efficient Reinforcement Learning for LLMs in Heterogeneous Environments
- Distributed Prioritized Experience Replay
- Reverb: A Framework For Experience Replay
- SEED RL: Scalable and Efficient Deep-RL with Accelerated Central Inference
- EARL: Efficient Agentic Reinforcement Learning Systems for Large Language Models
- DeepSpeed-Chat: Easy, Fast and Affordable RLHF Training of ChatGPT-like Models at All Scales
- DAPO: An Open-Source LLM Reinforcement Learning System at Scale
- SRL: Scaling Distributed Reinforcement Learning to Over Ten Thousand Cores
- Playing Atari with Deep Reinforcement Learning
- Prioritized Experience Replay
- Proximal Policy Optimization Algorithms
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
Related papers
- iScheduler: Reinforcement Learning-Driven Continual Optimization for Large-Scale Resource Investment Problems
- SAMM: Sharded Automated Market Maker
- InferScale: GPU-Native KV Injection for Personalized LLM Serving
- Vigil: Accountable Liveness against Selective Silence
- Steelhead: Interleaving Partially Synchronous and Asynchronous Commit Rules on a Shared DAG
- Pushing CPU Speech Synthesis to the Wall: Extreme Inference Tuning under Serverless Architecture and Billing