Good Scores, Bad Data: A Metric for Multimodal Coherence
For researchers developing multimodal AI systems, MCS provides a lightweight, annotation-free metric to diagnose fusion quality beyond task accuracy.
The paper introduces the Multimodal Coherence Score (MCS), a metric that evaluates fusion quality in multimodal AI systems independently of downstream task accuracy. MCS achieves higher sensitivity than task accuracy alone (Spearman rho = 0.093 vs. 0.071) and requires no human annotation.
Multimodal AI systems are evaluated by downstream task accuracy, but high accuracy does not mean the underlying data is coherent. A model can score well on Visual Question Answering (VQA) while its inputs contradict each other. We introduce the Multimodal Coherence Score (MCS), a metric that evaluates fusion quality independent of any downstream model. MCS decomposes coherence into four dimensions, identity, spatial, semantic, and decision, with weights learned via Nelder-Mead optimization. We evaluate on 1,000 Visual Genome images using DETR, CLIP, and ViLT, and validate on 150 COCO images with no retraining. Across three fusion architectures, MCS discriminates quality with higher sensitivity than task accuracy alone (Spearman rho = 0.093 vs. 0.071). Perturbation experiments confirm each dimension responds independently to its failure mode with zero cross-talk. MCS is lightweight, requires no human annotation, and tells you not just that something broke, but what broke.