AI Paper Atlas

Research map / Architectures and efficiency / Mixture of experts

MC#: Mixture Compressor for Mixture-of-Experts Large Models

Wei Huang, Yue Liao, Yukang Chen, Jianhui Liu, Haoru Tan, Si Liu, Shiming Zhang, Shuicheng Yan, Xiaojuan Qi

TPAMI 2026 · Architectures and efficiency › Mixture of experts · approach: Expert routing and specialization · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use PMQ mixed-precision expert bit allocation instead of uniform 2-bit (ablation)GPTQ on Mixtral 8x7b · eight zero-shot benchmarksaverage accuracy (%): 42.67 -> 63.25Improved
use PMQ instead of BSP at similar bits (ablation)Mixtral 8x7b · eight zero-shot benchmarksaverage accuracy (%): 49.07 -> 67.50Improved
use PMQ instead of Hessian allocation at 1.57-bit (ablation)Mixtral 8x7b · eight zero-shot benchmarksaverage accuracy (%): 45.91 -> 54.49Improved
use PMQ instead of Hessian allocation at 2.54-bit (ablation)Mixtral 8x7b · eight zero-shot benchmarksaverage accuracy (%): 67.18 -> 67.50No clear effect
use PMQ instead of uniform 2-bit (ablation)DeepSeek-VL2-T · five multimodal benchmarksaverage score (%): 1.41 -> 59.45Improved
use PMQ instead of Hessian at 2.58-bit (ablation)DeepSeek-VL2-S · five multimodal benchmarksaverage score (%): 61.79 -> 63.66Improved
add OTP learned pruning to PMQ (ablation)DeepSeek-VL2-S 2.07-bit · five multimodal benchmarksaverage score (%): 62.40 -> 60.92Hurt
learned OTP pruning instead of random pruning (ablation)PMQ DeepSeek-VL2-S · five multimodal benchmarksaverage score (%): 52.69 -> 60.92Improved
learned OTP instead of rule-based ODP pruning (ablation)PMQ Mixtral 8x7b · WikiText2perplexity (lower better): 6.62 -> 6.45Improved
add OTP to PMQ at 2.05-bit (ablation)Mixtral 8x7b · GSM8Kaccuracy (%): 19.97 -> 20.91No clear effect
add OTP to PMQ at 2.54-bit (ablation)Mixtral 8x7b · GSM8Kaccuracy (%): 37.67 -> 36.44Hurt
use PMQ+OTP at 2.05-bit instead of PMQ alone (ablation)Mixtral 8x7b · eight zero-shot benchmarksaverage accuracy (%): 63.25 -> 62.68Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy