AI Paper Atlas

Research map / Architectures and efficiency / Mixture of experts

Symphony-MoE: Harmonizing Disparate Pre-trained Models into a Coherent Mixture-of-Experts

Qi Wang, Hanyang Peng, Yue Yu

AAAI 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
add activation-based functional alignment (neuron permutation) (ablation)Symphony-MoE 1.5Bx4 · ID averageaverage score: 33.94 -> 44.07Improved
add activation-based functional alignment (ablation)Symphony-MoE 1.5Bx4 · MedCQA OODaccuracy (%): 26.92 -> 35.26Improved
align all neurons instead of only 80% (ablation)Symphony-MoE 1.5Bx4 · ID averageaverage score: 43.57 -> 44.07No clear effect
use SLERP instead of linear averaging for attention (ablation)Symphony-MoE 1.5Bx4 · ID averageaverage score: 37.28 -> 44.07Improved
use selective linear (MergeKit) embedding merge instead of naive averaging (ablation)Symphony-MoE 1.5Bx4 · ID averageaverage score: 40.19 -> 44.07Improved
sample calibration data from all expert domains instead of only general expert (ablation)Symphony-MoE 1.5Bx4 · MedCQA OODaccuracy (%): 29.44 -> 35.26Improved
Symphony-MoE vs Drop-UpcyclingMoE upcycling 0.5Bx4 · ID averageaverage score: 29.95 -> 31.77Improved
Symphony-MoE vs Drop-UpcyclingMoE upcycling 1.5Bx4 · MedCQA OODaccuracy (%): 32.90 -> 35.26Improved
Symphony-MoE vs code expert on HumanEvalSymphony-MoE 1.5Bx4 · HumanEvalpass@1: 43.41 -> 42.39Hurt
use general-purpose M1 as anchor instead of other experts (ablation)Symphony-MoE 1.5Bx4 · ID and OOD averageaverage scoreImproved
increase number of experts from 1 to 4 (ablation)Symphony-MoE 1.5Bx4 · ID and OODaverage scoreImproved

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