MSLA is the first physically deployable attack that uses adversarial lighting to break semantic alignment in VLMs such as CLIP, LLaVA, and BLIP, causing classification failures and hallucinations in real scenes.
Title resolution pending
24 Pith papers cite this work. Polarity classification is still indexing.
24
Pith papers citing it
citation-role summary
background 2
dataset 1
citation-polarity summary
representative citing papers
CrossMPI steers both visual and textual interpretations in LVLMs through image-only perturbations by optimizing in hidden-state space at selected middle layers with distance-based budget allocation.
AdaFocus achieves better accuracy on long-video benchmarks with roughly 33 times fewer visual tokens by combining query-aware adaptive sampling and zero-cache disk-based refinement.
In a controlled arithmetic-grammar program synthesis environment, diverse sampling across semantic and syntactic spaces yields robust density generalization while support generalization for novel syntax remains poor, with performance falling over 30 percent and compute scaling following a strictly 1
CGC improves fine-grained multi-image understanding in MLLMs by constructing contrastive training instances from existing single-image annotations and adding a rule-based spatial reward, achieving SOTA on MIG-Bench and VLM2-Bench with transfer gains to other multimodal tasks.
ONOTE is a multi-format benchmark that applies a deterministic pipeline to expose a disconnect between perceptual accuracy and music-theoretic comprehension in leading omnimodal AI models.
IAD-Unify unifies industrial anomaly segmentation, region-grounded language understanding, and mask-guided generation in one framework using DINOv2 token injection into Qwen3.5, supported by the new Anomaly-56K dataset of 59,916 images.
VISOR is a unified agentic VRAG framework with Evidence Space structuring, visual action evaluation/correction, and dynamic sliding-window trajectories trained via GRPO-based RL that achieves SOTA performance on long-horizon visual reasoning benchmarks.
Medical MLLMs degrade on image classification due to four failure modes in visual representation quality, connector projection fidelity, LLM comprehension, and semantic mapping alignment, quantified by feature probing on 14 models across 3 datasets.
A wrinkle-field perturbation method creates photorealistic non-rigid image changes that degrade state-of-the-art VLMs on image captioning and VQA more effectively than prior baselines.
A latent denoising objective with saliency-aware corruption and contrastive distillation improves visual alignment and corruption robustness in large multimodal models.
ST-π structures VLA models by having a spatiotemporal VLM produce causally ordered chunk-level prompts that guide a dual-generator action expert to jointly handle spatial and temporal control in robotic manipulation.
Q-Gate dynamically routes keyframe selection in long videos via query-modulated gating across visual grounding, global matching, and contextual alignment experts to improve MLLM performance.
HTDC mitigates hallucinations in LVLMs by triggering calibration only at hesitation-prone decoding steps via contrasts with visual-nullification and semantic-nullification probes.
GUIDE unrolls multi-granularity geometric priors layer-wise into early MLLM layers with gating to improve spatial reasoning and perception.
STEAR reduces spatial and temporal hallucinations in Video-LLMs via layer-aware evidence intervention from middle decoder layers in a single-encode pass.
A diffusion model with dynamic modality gating and cross-modal mutual learning restores missing features in VLMs bi-directionally while preserving the original model's generalization.
LaV-CoT introduces a multi-stage visual CoT pipeline and GRPO training with language-consistency rewards, delivering up to 9.5% accuracy gains on multilingual VQA benchmarks over similar-sized open models.
SKG-VLA models each complaint as a structured scene via a Scene Knowledge Graph to improve policy-grounded multimodal reasoning and decision accuracy.
SAKE is an agentic framework for GMNER that uses uncertainty-based self-awareness and reinforcement learning to balance internal knowledge exploitation with adaptive external exploration.
SpatialImaginer integrates visual imagination with textual chain-of-thought to improve spatial reasoning robustness in multimodal large language models.
MESA reduces hallucinations in LVLMs via controlled selective latent intervention that preserves the original token distribution.
Search-R3 trains LLMs to output search embeddings as a direct product of step-by-step reasoning via supervised pre-training and a specialized RL environment that avoids full corpus re-encoding.
citing papers explorer
-
Challenging Vision-Language Models with Physically Deployable Multimodal Semantic Lighting Attacks
MSLA is the first physically deployable attack that uses adversarial lighting to break semantic alignment in VLMs such as CLIP, LLaVA, and BLIP, causing classification failures and hallucinations in real scenes.
-
A Cross-Modal Prompt Injection Attack against Large Vision-Language Models with Image-Only Perturbation
CrossMPI steers both visual and textual interpretations in LVLMs through image-only perturbations by optimizing in hidden-state space at selected middle layers with distance-based budget allocation.
-
AdaFocus: Adaptive Relevance-Diversity Sampling with Zero-Cache Look-back for Efficient Long Video Understanding
AdaFocus achieves better accuracy on long-video benchmarks with roughly 33 times fewer visual tokens by combining query-aware adaptive sampling and zero-cache disk-based refinement.
-
Beyond the Training Distribution: Mapping Generalization Boundaries in Neural Program Synthesis
In a controlled arithmetic-grammar program synthesis environment, diverse sampling across semantic and syntactic spaces yields robust density generalization while support generalization for novel syntax remains poor, with performance falling over 30 percent and compute scaling following a strictly 1
-
CGC: Compositional Grounded Contrast for Fine-Grained Multi-Image Understanding
CGC improves fine-grained multi-image understanding in MLLMs by constructing contrastive training instances from existing single-image annotations and adding a rule-based spatial reward, achieving SOTA on MIG-Bench and VLM2-Bench with transfer gains to other multimodal tasks.
-
ONOTE: Benchmarking Omnimodal Notation Processing for Expert-level Music Intelligence
ONOTE is a multi-format benchmark that applies a deterministic pipeline to expose a disconnect between perceptual accuracy and music-theoretic comprehension in leading omnimodal AI models.
-
IAD-Unify: A Region-Grounded Unified Model for Industrial Anomaly Segmentation, Understanding, and Generation
IAD-Unify unifies industrial anomaly segmentation, region-grounded language understanding, and mask-guided generation in one framework using DINOv2 token injection into Qwen3.5, supported by the new Anomaly-56K dataset of 59,916 images.
-
VISOR: Agentic Visual Retrieval-Augmented Generation via Iterative Search and Over-horizon Reasoning
VISOR is a unified agentic VRAG framework with Evidence Space structuring, visual action evaluation/correction, and dynamic sliding-window trajectories trained via GRPO-based RL that achieves SOTA performance on long-horizon visual reasoning benchmarks.
-
Lost in the Hype: Revealing and Dissecting the Performance Degradation of Medical Multimodal Large Language Models in Image Classification
Medical MLLMs degrade on image classification due to four failure modes in visual representation quality, connector projection fidelity, LLM comprehension, and semantic mapping alignment, quantified by feature probing on 14 models across 3 datasets.
-
When Surfaces Lie: Exploiting Wrinkle-Induced Attention Shift to Attack Vision-Language Models
A wrinkle-field perturbation method creates photorealistic non-rigid image changes that degrade state-of-the-art VLMs on image captioning and VQA more effectively than prior baselines.
-
Latent Denoising Improves Visual Alignment in Large Multimodal Models
A latent denoising objective with saliency-aware corruption and contrastive distillation improves visual alignment and corruption robustness in large multimodal models.
-
ST-$\pi$: Structured SpatioTemporal VLA for Robotic Manipulation
ST-π structures VLA models by having a spatiotemporal VLM produce causally ordered chunk-level prompts that guide a dual-generator action expert to jointly handle spatial and temporal control in robotic manipulation.
-
Where to Focus: Query-Modulated Multimodal Keyframe Selection for Long Video Understanding
Q-Gate dynamically routes keyframe selection in long videos via query-modulated gating across visual grounding, global matching, and contextual alignment experts to improve MLLM performance.
-
HTDC: Hesitation-Triggered Differential Calibration for Mitigating Hallucination in Large Vision-Language Models
HTDC mitigates hallucinations in LVLMs by triggering calibration only at hesitation-prone decoding steps via contrasts with visual-nullification and semantic-nullification probes.
-
Let Geometry GUIDE: Layer-wise Unrolling of Geometric Priors in Multimodal LLMs
GUIDE unrolls multi-granularity geometric priors layer-wise into early MLLM layers with gating to improve spatial reasoning and perception.
-
STEAR: Layer-Aware Spatiotemporal Evidence Intervention for Hallucination Mitigation in Video Large Language Models
STEAR reduces spatial and temporal hallucinations in Video-LLMs via layer-aware evidence intervention from middle decoder layers in a single-encode pass.
-
Enhancing Foundation VLM Robustness to Missing Modality: Scalable Diffusion for Bi-directional Feature Restoration
A diffusion model with dynamic modality gating and cross-modal mutual learning restores missing features in VLMs bi-directionally while preserving the original model's generalization.
-
LaV-CoT: Language-Aware Visual CoT with Multi-Aspect Reward Optimization for Real-World Multilingual VQA
LaV-CoT introduces a multi-stage visual CoT pipeline and GRPO training with language-consistency rewards, delivering up to 9.5% accuracy gains on multilingual VQA benchmarks over similar-sized open models.
-
SKG-VLA: Scene Knowledge Graph Priors for Structured Scene Semantics and Multimodal Reasoning for Decision Making
SKG-VLA models each complaint as a structured scene via a Scene Knowledge Graph to improve policy-grounded multimodal reasoning and decision accuracy.
-
SAKE: Self-aware Knowledge Exploitation-Exploration for Grounded Multimodal Named Entity Recognition
SAKE is an agentic framework for GMNER that uses uncertainty-based self-awareness and reinforcement learning to balance internal knowledge exploitation with adaptive external exploration.
-
SpatialImaginer: Towards Adaptive Visual Imagination for Spatial Reasoning
SpatialImaginer integrates visual imagination with textual chain-of-thought to improve spatial reasoning robustness in multimodal large language models.
-
Mitigating Entangled Steering in Large Vision-Language Models for Hallucination Reduction
MESA reduces hallucinations in LVLMs via controlled selective latent intervention that preserves the original token distribution.
-
Search-R3: Unifying Reasoning and Embedding in Large Language Models
Search-R3 trains LLMs to output search embeddings as a direct product of step-by-step reasoning via supervised pre-training and a specialized RL environment that avoids full corpus re-encoding.
- Visualizing the Invisible: Generative Visual Grounding Empowers Universal EEG Understanding in MLLMs