EfficientRollout: System-Aware Self-Speculative Decoding for RL Rollouts

arXiv:2606.18967 · cs.LG · Submitted 2026-06-17 · Read on arXiv

cs.LG

Submitted: 2026-06-17

Updated: 2026-09-30

Comments: Accepted to NeurIPS 2026; Project Page: https://github.com/furiosa-ai/EfficientRollout

Code: https://github.com/furiosa-ai/EfficientRollout

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Speculative Decoding — 3× in this paper · explained in 2 episodes
• Adaptation — 2× in this paper · explained in 1 episode
• Verification — 2× in this paper · explained in 2 episodes
• Decode — 1× in this paper · explained in 2 episodes
• Online Adaptation — 1× in this paper · explained in 3 episodes
• Reinforcement Learning — 1× in this paper · explained in 4 episodes
• Reinforcement Learning (RL) — 1× in this paper · explained in 25 episodes

Sources

• MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention
• Accelerating Large Language Model Decoding with Speculative Sampling
• Jackpot: Optimal Budgeted Rejection Sampling for Extreme Actor-Policy Mismatch Reinforcement Learning
• Multi-Head Attention: Collaborate Instead of Concatenate
• The Entropy Mechanism of Reinforcement Learning for Reasoning Language Models
• The Llama 3 Herd of Models
• DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
• History Rhymes: Accelerating LLM Reinforcement Learning with RhymeRL
• Measuring Mathematical Problem Solving With the MATH Dataset
• Accelerating RL Post-Training Rollouts via System-Integrated Speculative Decoding
• Revisiting Entropy in Reinforcement Learning for Large Reasoning Models
• Beyond Next-Token Prediction: A Performance Characterization of Diffusion versus Autoregressive Language Models
• LLM Post-Training: A Deep Dive into Reasoning Large Language Models
• DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model
• SPEC-RL: Accelerating On-Policy Reinforcement Learning with Speculative Rollouts
• Understanding R1-Zero-Like Training: A Critical Perspective
• Proximal Policy Optimization Algorithms
• DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
• Fast Transformer Decoding: One Write-Head is All You Need
• Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks