AI Paper Atlas

Research map / Multi-agent systems and games / LLM agents

AIM: Adaptive Interaction in Multi-Agent Debate for Multimodal LLM Inference

Wei Fan, JinYi Yoon, Bo Ji

NeurIPS 2026 · Multi-agent systems and games › LLM agents · approach: Self-evolving language agent learning · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes AIM, an adaptive multi-agent debate framework for multimodal LLMs that selects which interaction modality agents use to exchange evidence, instead of fixed text- or graph-based interaction.

Problem

Multi-Agent Debate for multimodal reasoning depends on how agents interact, but existing methods use fixed text- or graph-based interaction that fails to convey fine-grained multimodal evidence, hurting accuracy and inflating token cost.

Approach

Treats the interaction modality (which modality agents use to exchange evidence) as an adaptive design choice in multi-agent debate, rather than a fixed text or graph channel.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
exchange task-dependent evidence (image RoI, audio segment, video clip) instead of text-only interaction (ablation)multimodal MAD · GQAaccuracy: 68.9 -> 71.2Improved
combine text with task-dependent interaction (T+X) instead of X alone (ablation)multimodal MAD · GQAaccuracy: 71.2 -> 77.6Improved
add graph modality to T+X (all three modalities) (ablation)multimodal MAD · GQAaccuracy: 77.6 -> 77.5No clear effect
add graph modality to T+X (all three modalities) (ablation)multimodal MAD · DCASEaccuracy: 50.9 -> 49.8Hurt
use AIM adaptive per-instance routing instead of the best fixed combination (T+G+X) on OKVQA (ablation)multimodal MAD · OKVQAaccuracy: 93.5 -> 93.8 (tokens 9,773 -> 6,775)Improved
use AIM adaptive routing instead of fixed T+G+X (ablation)multimodal MAD · VCRaccuracy: 79.6 -> 84.2Improved
use AIM routing instead of best fixed combination (ablation)multimodal MAD · IntentQAaccuracy: 71.2 -> 75.4Improved
route with question-type features only instead of all four feature groups (ablation)AIM router · eight benchmarks (average)accuracy: AIM improves average accuracy by 1.16% over question-type routingHurt
use cost-aware heuristic router only (no correctness features) (ablation)AIM router · eight benchmarks (average)accuracy: 2.51% lower average accuracyHurt
use any single-group adaptive router instead of fixed T+X (ablation)multimodal MAD · eight benchmarks (average)accuracy: +1.0% on averageImproved
use richer modality combinations rather than single modalitymultimodal MAD · GQAFaithScore: 0.50 (G) / 0.61 (T) -> 0.72 (AIM)Improved
AIM vs T+X in FaithScoremultimodal MAD · VQA-v2FaithScore: 0.78 -> 0.77No clear effect

Experimental setup

Metrics: accuracy token cost

Baselines: text-based multi-agent debate graph-based multi-agent debate

Builds on: multi-agent debate

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