LEMON: Local Explanations via Modality-aware OptimizatioN

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Qin, Yu, Sloan, Phillip, Santos-Rodriguez, Raul, Mirmehdi, Majid, Filho, Telmo de Menezes e Silva
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915770202587136
author Qin, Yu
Sloan, Phillip
Santos-Rodriguez, Raul
Mirmehdi, Majid
Filho, Telmo de Menezes e Silva
author_facet Qin, Yu
Sloan, Phillip
Santos-Rodriguez, Raul
Mirmehdi, Majid
Filho, Telmo de Menezes e Silva
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2602_02786
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle LEMON: Local Explanations via Modality-aware OptimizatioN
Qin, Yu
Sloan, Phillip
Santos-Rodriguez, Raul
Mirmehdi, Majid
Filho, Telmo de Menezes e Silva
Machine Learning
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.
title LEMON: Local Explanations via Modality-aware OptimizatioN
topic Machine Learning
url https://arxiv.org/abs/2602.02786