A single input-agnostic thermal-airflow perturbation, optimized on one surrogate CLIP model, transfers to five CLIP backbones and six VLMs, degrading scene classification by up to 38.2%.
Advances in neural information processing systems36, 49250–49267 (2023)
14 Pith papers cite this work. Polarity classification is still indexing.
14
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
years
2026 14roles
background 2polarities
background 2representative citing papers
DO-Bench is a controlled benchmark that attributes VLM object hallucination errors to textual prior pressure, perceptual limits, or their interaction via two diagnostic dimensions and metrics.
DeP mitigates MLLM hallucinations by dynamically perturbing text prompts to identify and reinforce stable visual evidence regions while counteracting language prior biases using attention variance and logit statistics.
RESP uses reference-guided sequential prompting with VLMs to improve frame-level and video-level visual glitch detection in games by establishing per-video baselines.
PromptEvolver recovers high-fidelity natural language prompts for given images by evolving them via genetic algorithm guided by a vision-language model, outperforming prior methods on benchmarks.
An agentic mRAG framework uses GRPO-trained visual reranking and active rejection to verify retrieved candidate entities, achieving state-of-the-art on three KB-VQA benchmarks.
StochasT uses stochastic clustering of language tasks into varying turn depths for the same image to improve LVLMs on both single-turn and multi-turn scenarios without discarding data.
ReShift is a reasoning-level backdoor framework for VLMs that uses poisoned data construction and joint optimization to shift CoT trajectories on trigger while preserving surface coherence.
MV-GEL uses a learned view selector and a fine-tuned vision-language segmentation model to localize text-described faces and edges on 3D meshes.
A two-stage framework enables multimodal LLMs to learn shared latent representations from pairwise modality data and achieve cross-modal generation when incorporating new modalities.
ReflectCAP distills model-specific hallucination and oversight patterns into Structured Reflection Notes that steer LVLMs toward more factual and complete image captions, reaching the Pareto frontier on factuality-coverage trade-offs.
Introduces ML-FOP-SOAP optimizer using Fisher-Orthogonal Projection and hierarchical folding to mitigate modality competition in multimodal autoregressive training, reporting gains over AdamW on Janus and Emu3.
OpenSpatial supplies a principled open-source data engine and 3-million-sample dataset that raises spatial-reasoning model performance by an average of 19 percent on benchmarks.
VLMs fine-tuned on a consistency-probed Visual-Idk dataset via SFT and preference optimization raise truthful rate from 57.9% to 67.3% and show internal evidence of genuine boundary recognition.
citing papers explorer
-
AirflowAttack: Thermal-Airflow Adversarial Perturbations against Infrared Remote-Sensing Vision-Language Models
A single input-agnostic thermal-airflow perturbation, optimized on one surrogate CLIP model, transfers to five CLIP backbones and six VLMs, degrading scene classification by up to 38.2%.
-
DO-Bench: An Attributable Benchmark for Diagnosing Object Hallucination in Vision-Language Models
DO-Bench is a controlled benchmark that attributes VLM object hallucination errors to textual prior pressure, perceptual limits, or their interaction via two diagnostic dimensions and metrics.
-
Decoding by Perturbation: Mitigating MLLM Hallucinations via Dynamic Textual Perturbation
DeP mitigates MLLM hallucinations by dynamically perturbing text prompts to identify and reinforce stable visual evidence regions while counteracting language prior biases using attention variance and logit statistics.
-
RESP: Reference-guided Sequential Prompting for Visual Glitch Detection in Video Games
RESP uses reference-guided sequential prompting with VLMs to improve frame-level and video-level visual glitch detection in games by establishing per-video baselines.
-
PromptEvolver: Prompt Inversion through Evolutionary Optimization in Natural-Language Space
PromptEvolver recovers high-fidelity natural language prompts for given images by evolving them via genetic algorithm guided by a vision-language model, outperforming prior methods on benchmarks.
-
MMAgent-R$^2$: Learning to Rerank and Reject for Agentic mRAG
An agentic mRAG framework uses GRPO-trained visual reranking and active rejection to verify retrieved candidate entities, achieving state-of-the-art on three KB-VQA benchmarks.
-
StochasT: Learning with Stochastic Turn Depth for Visual Instruction Tuning
StochasT uses stochastic clustering of language tasks into varying turn depths for the same image to improve LVLMs on both single-turn and multi-turn scenarios without discarding data.
-
ReShift: Aha-Moment-Driven Reasoning-Level Backdoor Attacks on Vision-Language Models
ReShift is a reasoning-level backdoor framework for VLMs that uses poisoned data construction and joint optimization to shift CoT trajectories on trigger while preserving surface coherence.
-
MV-GEL: Language-Driven Multi-View Geometric Entity Localization on Meshes
MV-GEL uses a learned view selector and a fine-tuned vision-language segmentation model to localize text-described faces and edges on 3D meshes.
-
Multimodal LLMs under Pairwise Modalities
A two-stage framework enables multimodal LLMs to learn shared latent representations from pairwise modality data and achieve cross-modal generation when incorporating new modalities.
-
ReflectCAP: Detailed Image Captioning with Reflective Memory
ReflectCAP distills model-specific hallucination and oversight patterns into Structured Reflection Notes that steer LVLMs toward more factual and complete image captions, reaching the Pareto frontier on factuality-coverage trade-offs.
-
Second-Order Multi-Level Variance Correction for Modality Competition in Multimodal Models
Introduces ML-FOP-SOAP optimizer using Fisher-Orthogonal Projection and hierarchical folding to mitigate modality competition in multimodal autoregressive training, reporting gains over AdamW on Janus and Emu3.
-
OpenSpatial: A Principled Data Engine for Empowering Spatial Intelligence
OpenSpatial supplies a principled open-source data engine and 3-million-sample dataset that raises spatial-reasoning model performance by an average of 19 percent on benchmarks.
-
Delineating Knowledge Boundaries for Honest Large Vision-Language Models
VLMs fine-tuned on a consistency-probed Visual-Idk dataset via SFT and preference optimization raise truthful rate from 57.9% to 67.3% and show internal evidence of genuine boundary recognition.