LEMON: Local Explanations via Modality-aware OptimizatioN

TL;DR

LEMON is a model-agnostic framework that provides efficient, unified local explanations for multimodal models by disentangling modality and feature contributions, applicable to vision-language and clinical tasks.

Contribution

LEMON introduces a novel modality-aware surrogate with group sparsity for faithful, scalable explanations of multimodal predictions, outperforming existing methods in efficiency.

Findings

LEMON reduces black-box evaluations by up to 67 times.

LEMON maintains competitive faithfulness in explanations.

LEMON is applicable to diverse multimodal tasks like vision-language and clinical prediction.

Abstract

Multimodal models are ubiquitous, yet existing explainability methods are often single-modal, architecture-dependent, or too computationally expensive to run at scale. We introduce LEMON (Local Explanations via Modality-aware OptimizatioN), a model-agnostic framework for local explanations of multimodal predictions. LEMON fits a single modality-aware surrogate with group-structured sparsity to produce unified explanations that disentangle modality-level contributions and feature-level attributions. The approach treats the predictor as a black box and is computationally efficient, requiring relatively few forward passes while remaining faithful under repeated perturbations. We evaluate LEMON on vision-language question answering and a clinical prediction task with image, text, and tabular inputs, comparing against representative multimodal baselines. Across backbones, LEMON achieves competitive deletion-based faithfulness while reducing black-box evaluations by 35-67 times and runtime by 2-8 times compared to strong multimodal baselines.