RADAR: Accelerate Large Language Model Inference With RL-Based Dynamic Draft Trees

arXiv:2512.14069 · cs.AI · Submitted 2025-12-16 · Read on arXiv

cs.AI

Submitted: 2025-12-16

Updated: 2026-08-26

Comments: EMNLP 2026 Findings Paper

Code: https://github.com/minaduki-sora/RADAR

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

Terminology

• Reinforcement Learning — 11× in this paper · explained in 4 episodes
• Verification — 8× in this paper · explained in 2 episodes
• Markov Decision Process — 4× in this paper
• Large Language Models (LLMs) — 3× in this paper · explained in 6 episodes
• Markov Decision Process (MDP) — 3× in this paper · explained in 11 episodes
• Reward Function — 3× in this paper · explained in 2 episodes
• Distillation — 2× in this paper · explained in 2 episodes
• Pruning — 2× in this paper · explained in 4 episodes
• Quantization — 2× in this paper · explained in 12 episodes
• Transformer — 2× in this paper · explained in 2 episodes
• Action Space — 1× in this paper · explained in 1 episode
• Benchmark — 1× in this paper · explained in 2 episodes
• Computational overhead — 1× in this paper · explained in 2 episodes
• Core — 1× in this paper · explained in 2 episodes
• Deep Reinforcement Learning — 1× in this paper · explained in 2 episodes
• Elias — 1× in this paper · explained in 2 episodes
• Fine-Tuning — 1× in this paper · explained in 5 episodes
• Knowledge Distillation — 1× in this paper · explained in 11 episodes
• LLM-as-a-Judge — 1× in this paper · explained in 3 episodes
• Loss Function — 1× in this paper · explained in 2 episodes
• On-Policy Distillation — 1× in this paper · explained in 3 episodes
• Online Reinforcement Learning — 1× in this paper · explained in 2 episodes
• Speculative Decoding — 1× in this paper · explained in 2 episodes
• Verifier — 1× in this paper · explained in 2 episodes

Sources

• GPT-4 Technical Report
• DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
• Efficient Prompting via Dynamic In-Context Learning
• GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers
• Accelerating Large Language Model Decoding with Speculative Sampling
• EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty
• EAGLE-2: Faster Inference of Language Models with Dynamic Draft Trees
• Learning Harmonized Representations for Speculative Sampling
• Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads
• Hydra: Sequentially-Dependent Draft Heads for Medusa Decoding
• EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Test
• Training Verifiers to Solve Math Word Problems
• Program Synthesis with Large Language Models

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems