AcceptMoE: Commitment-Weighted Self-Sizing Verifier Expert Sets for Efficient MoE Speculative Decoding
Shuang Liang, Hao Mark Chen, Zhiwen Mo, Qianzhou Wang, Guoyu Li, Lingxiao Ma, Wayne Luk
cs.LG, cs.CL, cs.DC
Submitted: 2026-08-04
Comments: 10 pages, 5 figures
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- gpt-oss-120b & gpt-oss-20b Model Card
- DA-MoE: Towards Dynamic Expert Allocation for Mixture-of-Experts Models
- Program Synthesis with Large Language Models
- Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads
- Dynamic Expert Sharing: Decoupling Memory from Parallelism in Mixture-of-Experts Diffusion LLMs
- SP-MoE: Speculative Decoding and Prefetching for Accelerating MoE-based Model Inference
- Evaluating Large Language Models Trained on Code
- Task-Specific Expert Pruning for Sparse Mixture-of-Experts
- Training Verifiers to Solve Math Word Problems
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- Mixtral of Experts
- MoE-Spec: Expert Budgeting for Efficient Speculative Decoding
- Opportunistic Expert Activation: Batch-Aware Expert Routing for Faster Decode Without Retraining
- Making Every Verified Token Count: Adaptive Verification for MoE Speculative Decoding
- Utility-Driven Speculative Decoding for Mixture-of-Experts
- XShare: Collaborative in-Batch Expert Sharing for Faster MoE Inference
- Specter: Fast MoE Inference via Quantization-Enabled Multi-Step Speculative Offloading
- Less Experts, Faster Decoding: Cost-Aware Speculative Decoding for Mixture-of-Experts
- Qwen3 Technical Report
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