AI Paper Atlas

Research map / Architectures and efficiency / Mixture of experts

On the Representation Collapse of Sparse Mixture of Experts

Zewen Chi, Li Dong, Shaohan Huang, Damai Dai, Shuming Ma, Barun Patra, Saksham Singhal, Payal Bajaj, Xia Song, Xian-Ling Mao, Heyan Huang, Furu Wei

NeurIPS 2022 · Architectures and efficiency › Mixture of experts · approach: Expert routing and specialization · 20 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
replace dot-product routing with low-dim L2-normalized routing and learnable temperature (X-MoE)SMoE (softmax gating) · XTREME averageavg score: 63.8 -> 65.3Improved
replace dot-product routing with X-MoE routingSMoE (sigmoid gating) · XTREME averageavg score: 64.1 -> 65.0Improved
use X-MoE routingSMoE · WMT-10 x->enavg BLEU: 29.7 -> 30.2Improved
add sparse MoE layer to dense modelDense Transformer · XTREME averageavg score: 61.4 -> 63.8Improved
X-MoE routing vs baseline on upstream MLMSMoE · mC4 MLM validationperplexityImproved
add dimension reduction, L2 norm and frozen routing (each component) (ablation)X-MoE sigmoid · XNLI and MLQAfine-tuning scoreImproved
routing dimension N/2 vs other sizes (N/4..4N) (ablation)X-MoE sigmoid · XNLI / MLQAscoreMixed
add load balancing loss during fine-tuning (ablation)X-MoE · XNLI and MLQAscoreNo clear effect

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