Research map / Architectures and efficiency / Mixture of experts
Route-Consistent Adaptation for Stable Quantization of Mixture-of-Experts Models with Theoretical Guarantees
Longteng Zhang, Sen Wu, Qiang Wang, Shaohuai Shi, Xiaowen Chu
Read the paperCheck your idea against 137,217 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes a unified quantization framework for Mixture-of-Experts models that jointly handles expert quantization error and routing-induced instability, with theoretical guarantees.
Problem
Quantizing MoE models is hard because expert quantization error perturbs expert outputs and sparse routing amplifies these perturbations into discrete expert selection changes, causing progressive output drift not captured by dense-model quantization analyses.
Approach
Jointly handles expert quantization error and routing instability in one framework (route-consistent adaptation), with theoretical analysis of the coupled drift; the abstract is truncated so specifics are not stated.
Closest papers by meaning
- Quant Experts: Token-aware Adaptive Error Reconstruction with Mixture of Experts for Large Vision-Language Models Quantization (CVPR 2026 · similarity 0.89)
- MoEQuant: Enhancing Quantization for Mixture-of-Experts Large Language Models via Expert-Balanced Sampling and Affinity Guidance (ICML 2025 · similarity 0.89)
- Dense Backpropagation Improves Training for Sparse Mixture-of-Experts (NeurIPS 2025 · similarity 0.88)
- Mixture Compressor for Mixture-of-Experts LLMs Gains More (ICLR 2025 · similarity 0.88)
- Tight Clusters Make Specialized Experts (ICLR 2025 · similarity 0.88)
- A Provably Effective Method for Pruning Experts in Fine-tuned Sparse Mixture-of-Experts (ICML 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
