Cover Image

More Images, More Problems? A Controlled Analysis of VLM Failure Modes

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Das, Anurag, Bulat, Adrian, Baldrati, Alberto, Metaxas, Ioannis Maniadis, Schiele, Bernt, Tzimiropoulos, Georgios, Martinez, Brais
Format:	Preprint
Published:	2026
Subjects:	Computer Vision and Pattern Recognition
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

VISion On Request: Enhanced VLLM efficiency with sparse, dynamically selected, vision-language interactions
by: Bulat, Adrian, et al.
Published: (2026)

Aligned Unsupervised Pretraining of Object Detectors with Self-training
by: Metaxas, Ioannis Maniadis, et al.
Published: (2023)

VladVA: Discriminative Fine-tuning of LVLMs
by: Ouali, Yassine, et al.
Published: (2024)

Efficient Unsupervised Visual Representation Learning with Explicit Cluster Balancing
by: Metaxas, Ioannis Maniadis, et al.
Published: (2024)

CLIP-DPO: Vision-Language Models as a Source of Preference for Fixing Hallucinations in LVLMs
by: Ouali, Yassine, et al.
Published: (2024)

Hierarchical Image Tokenization for Multi-Scale Image Super Resolution
by: Hadji, Isma, et al.
Published: (2026)

You Only Need One Step: Fast Super-Resolution with Stable Diffusion via Scale Distillation
by: Noroozi, Mehdi, et al.
Published: (2024)

Knowledge Distillation Meets Open-Set Semi-Supervised Learning
by: Yang, Jing, et al.
Published: (2022)

Multi-scale Image Super Resolution with a Single Auto-Regressive Model
by: Sanchez, Enrique, et al.
Published: (2025)

FAM Diffusion: Frequency and Attention Modulation for High-Resolution Image Generation with Stable Diffusion
by: Yang, Haosen, et al.
Published: (2024)

FFF: Fixing Flawed Foundations in contrastive pre-training results in very strong Vision-Language models
by: Bulat, Adrian, et al.
Published: (2024)

Fwd2Bot: LVLM Visual Token Compression with Double Forward Bottleneck
by: Bulat, Adrian, et al.
Published: (2025)

Restore, Assess, Repeat: A Unified Framework for Iterative Image Restoration
by: Chen, I-Hsiang, et al.
Published: (2026)

VITAL: More Understandable Feature Visualization through Distribution Alignment and Relevant Information Flow
by: Gorgun, Ada, et al.
Published: (2025)

ClipTTT: CLIP-Guided Test-Time Training Helps LVLMs See Better
by: Nath, Mriganka, et al.
Published: (2026)

MTA-CLIP: Language-Guided Semantic Segmentation with Mask-Text Alignment
by: Das, Anurag, et al.
Published: (2024)

Vision-Free Retrieval: Rethinking Multimodal Search with Textual Scene Descriptions
by: Ntinou, Ioanna, et al.
Published: (2025)

CLIPCleaner: Cleaning Noisy Labels with CLIP
by: Feng, Chen, et al.
Published: (2024)

Do Instance Priors Help Weakly Supervised Semantic Segmentation?
by: Das, Anurag, et al.
Published: (2026)

Edge-SD-SR: Low Latency and Parameter Efficient On-device Super-Resolution with Stable Diffusion via Bidirectional Conditioning
by: Noroozi, Mehdi, et al.
Published: (2024)

SSR: An Efficient and Robust Framework for Learning with Unknown Label Noise
by: Feng, Chen, et al.
Published: (2021)

DWDN: Deep Wiener Deconvolution Network for Non-Blind Image Deblurring
by: Dong, Jiangxin, et al.
Published: (2021)

CemiFace: Center-based Semi-hard Synthetic Face Generation for Face Recognition
by: Sun, Zhonglin, et al.
Published: (2024)

MM2Latent: Text-to-facial image generation and editing in GANs with multimodal assistance
by: Meng, Debin, et al.
Published: (2024)

A Little More Like This: Text-to-Image Retrieval with Vision-Language Models Using Relevance Feedback
by: Khaertdinov, Bulat, et al.
Published: (2025)

DiffusionAct: Controllable Diffusion Autoencoder for One-shot Face Reenactment
by: Bounareli, Stella, et al.
Published: (2024)

LAFS: Landmark-based Facial Self-supervised Learning for Face Recognition
by: Sun, Zhonglin, et al.
Published: (2024)

OrCo: Towards Better Generalization via Orthogonality and Contrast for Few-Shot Class-Incremental Learning
by: Ahmed, Noor, et al.
Published: (2024)

One-shot Neural Face Reenactment via Finding Directions in GAN's Latent Space
by: Bounareli, Stella, et al.
Published: (2024)

Deconstructing the Failure of Ideal Noise Correction: A Three-Pillar Diagnosis
by: Feng, Chen, et al.
Published: (2026)

iSEARLE: Improving Textual Inversion for Zero-Shot Composed Image Retrieval
by: Agnolucci, Lorenzo, et al.
Published: (2024)

Multiscale Vision Transformers meet Bipartite Matching for efficient single-stage Action Localization
by: Ntinou, Ioanna, et al.
Published: (2023)

MeMSVD: Long-Range Temporal Structure Capturing Using Incremental SVD
by: Ntinou, Ioanna, et al.
Published: (2024)

Optimising for Interpretability: Convolutional Dynamic Alignment Networks
by: Böhle, Moritz, et al.
Published: (2021)

Towards Better Understanding Attribution Methods
by: Rao, Sukrut, et al.
Published: (2022)

Multimodal-Conditioned Latent Diffusion Models for Fashion Image Editing
by: Baldrati, Alberto, et al.
Published: (2024)

Training-Free Generation of Diverse and High-Fidelity Images via Prompt Semantic Space Optimization
by: Meng, Debin, et al.
Published: (2025)

Better Understanding Differences in Attribution Methods via Systematic Evaluations
by: Rao, Sukrut, et al.
Published: (2023)

Adversarial Training against Location-Optimized Adversarial Patches
by: Rao, Sukrut, et al.
Published: (2020)

B-cos Alignment for Inherently Interpretable CNNs and Vision Transformers
by: Böhle, Moritz, et al.
Published: (2023)