AI Paper Atlas

Research map / Large language models / Agents, code and math

MCCE: A Framework for Multi-LLM Collaborative Search in Discrete Spaces with Similarity-Filtered Preference Learning

Nian Ran, Zhongzheng Li, Yue Wang, Qingsong Ran, Xiaoyuan Zhang, Shikun Feng, Richard Allmendinger, Xiaoguang Zhao

ICML 2026 · Large language models › Agents, code and math · approach: LLMs for optimization and heuristics · 0 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 SFT with DPO for refining the local model (ablation)MCCE · 5-objective molecular designhypervolume: 0.709 -> 0.847Improved
replace RL with DPO for refining the local model (ablation)MCCE · 5-objective molecular designhypervolume: 0.709 -> 0.847Improved
add parameter training of the local model (DPO) to plain two-LLM collaboration (ablation)MCCE · 5-objective molecular designhypervolume: 0.695 -> 0.847Improved
add the frozen API LLM to a lone local Qwen2.5-7B-Instruct (ablation)MCCE · 5-objective molecular designhypervolume: 0.516 -> 0.847Improved
replace trainable local model with a second frozen API (GPT-4o + Gemini) (ablation)MCCE · 5-objective molecular designhypervolume: 0.847 -> 0.492Hurt
use only the relaxed similarity window I3 instead of nested windows (ablation)MCCE similarity filtering · 5-objective molecular designTop1 fitness: 4.35 -> 4.32No clear effect
use embedding-based similarity instead of fingerprint similarity (ablation)MCCE similarity filtering · 5-objective molecular designTop1 fitness: 4.35 -> 4.29Hurt
vary score quantile alpha from 0.30 to 0.20 or 0.40 (ablation)MCCE · 5-objective molecular designhypervolume: 0.847 -> 0.828 (0.20), 0.843 (0.40)No clear effect
use 50/16 API/local call split instead of 50/32 (ablation)MCCE · 5-objective molecular designhypervolume: 0.847 -> 0.734Hurt
use 50/50 API/local call split instead of 50/32 (ablation)MCCE · 5-objective molecular designTop1 fitness: 4.35 -> 4.31Hurt
update the local model less often (every 500 candidates instead of one round) (ablation)MCCE · 5-objective molecular designhypervolume: 0.847 -> 0.796Hurt
update every 200 candidates instead of default (ablation)MCCE · 5-objective molecular designhypervolume: 0.847 -> 0.856No 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