AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Securing Multimodal AI through Internal Information Decomposition

Jehyeok Yeon, Hyeonjeong Ha, Qiusi Zhan, Heng Ji

ICML 2026 · Vision-language and multimodal › Cross-modal learning · approach: Multimodal fusion with missing modalities · 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
remove Synergy feature from FlowVector (ablation)FlowGuard · VLSafeAUC: -0.125 on VLSafe, -0.110 on VLSU when removedImproved
include Visual Uniqueness (Uv) feature (ablation)FlowGuard · VizWiz (benign OOD)AUC: prevents a -0.150 dropImproved
use FlowVector relational features instead of raw final-layer embeddings with the same anomaly detector (ablation)FlowGuard · MM-SafetyBench UMK, LLaVA-1.5-7BASR: 52.8 -> 8.7Improved
monitor more decoding steps (k=1 to k=10) (ablation)FlowGuard · Detection AUCAUC: 0.942 -> 0.948, latency 1.3s -> 12.8sNo clear effect
monitor k=3 decoding steps instead of 1 (ablation)FlowGuard · Adversarial samplesevasion catch rate: 2.1% of evading samples caught at k=3Improved
use supervised MLP instead of Isolation Forest on FlowVectors (ablation)FlowGuard · Unseen attacksAUC: Isolation Forest AUC >= 0.88 on OODHurt
add FlowGuard to undefended model against co-optimized GCG+PGD attackLLaVA-1.5-7B · Co-optimized GCG and PGD attackASR: 86.4% -> 11.2%Improved
FlowGuard instead of perplexity or confidence-based detectionDetector · Benign inputsFPR: 2.4% vs 11.2% (perplexity), 14.5% (confidence)Improved
use FlowGuard on text-only attacks versus visual or cross-modal attacksFlowGuard · Text attacksASR: residual ASR about 13-14% text vs 6-9% visual/cross-modalMixed
generative verification defenses (CIDER, MirrorCheck) on low-quality imagesCIDER and MirrorCheck · VizWiz-VQAaccuracy: 56.2 -> 35.2 (CIDER), 56.2 -> 38.1 (MirrorCheck)Hurt

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