Cover Image

The Geometry of Robustness: Optimizing Loss Landscape Curvature and Feature Manifold Alignment for Robust Finetuning of Vision-Language Models

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Chopra, Shivang, Halbe, Shaunak, Huang, Chengyue, Maneechotesuwan, Brisa, Kira, Zsolt
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

FRAMES-VQA: Benchmarking Fine-Tuning Robustness across Multi-Modal Shifts in Visual Question Answering
by: Huang, Chengyue, et al.
Published: (2025)

Directional Gradient Projection for Robust Fine-Tuning of Foundation Models
by: Huang, Chengyue, et al.
Published: (2025)

Continual Adaptation of Vision Transformers for Federated Learning
by: Halbe, Shaunak, et al.
Published: (2023)

Mimicking or Reasoning: Rethinking Multi-Modal In-Context Learning in Vision-Language Models
by: Huang, Chengyue, et al.
Published: (2025)

Rethinking Weight Decay for Robust Fine-Tuning of Foundation Models
by: Tian, Junjiao, et al.
Published: (2024)

MedMoE: Modality-Specialized Mixture of Experts for Medical Vision-Language Understanding
by: Chopra, Shivang, et al.
Published: (2025)

Adaptive Memory Replay for Continual Learning
by: Smith, James Seale, et al.
Published: (2024)

MAPS: Preserving Vision-Language Representations via Module-Wise Proximity Scheduling for Better Vision-Language-Action Generalization
by: Huang, Chengyue, et al.
Published: (2025)

Grounding Descriptions in Images informs Zero-Shot Visual Recognition
by: Halbe, Shaunak, et al.
Published: (2024)

Anchor-based Robust Finetuning of Vision-Language Models
by: Han, Jinwei, et al.
Published: (2024)

VeRVE: Versatile Retrieval for Videos via Unified Embeddings
by: Halbe, Shaunak, et al.
Published: (2026)

GAMA: Geometry-Aware Manifold Alignment via Structured Adversarial Perturbations for Robust Domain Adaptation
by: Satou, Hana, et al.
Published: (2025)

AME: Aligned Manifold Entropy for Robust Vision-Language Distillation
by: Cao, Guiming, et al.
Published: (2025)

Approximate Nullspace Augmented Finetuning for Robust Vision Transformers
by: Liu, Haoyang, et al.
Published: (2024)

Sim2real Image Translation Enables Viewpoint-Robust Policies from Fixed-Camera Datasets
by: Coholich, Jeremiah, et al.
Published: (2026)

Are Classification Robustness and Explanation Robustness Really Strongly Correlated? An Analysis Through Input Loss Landscape
by: Chen, Tiejin, et al.
Published: (2024)

SAGE: Sink-Aware Grounded Decoding for Multimodal Hallucination Mitigation
by: Shukla, Tripti, et al.
Published: (2026)

Robust Hyperbolic Learning with Curvature-Aware Optimization
by: Bdeir, Ahmad, et al.
Published: (2024)

Finetune Like You Pretrain: Boosting Zero-shot Adversarial Robustness in Vision-language Models
by: Xing, Songlong, et al.
Published: (2026)

UMERegRobust - Universal Manifold Embedding Compatible Features for Robust Point Cloud Registration
by: Haitman, Yuval, et al.
Published: (2024)

Proxy Robustness in Vision Language Models is Effortlessly Transferable
by: Fu, Xiaowei, et al.
Published: (2026)

Transcending Domains through Text-to-Image Diffusion: A Source-Free Approach to Domain Adaptation
by: Chopra, Shivang, et al.
Published: (2023)

Overcoming the Pitfalls of Vision-Language Model Finetuning for OOD Generalization
by: Zang, Yuhang, et al.
Published: (2024)

Quality Text, Robust Vision: The Role of Language in Enhancing Visual Robustness of Vision-Language Models
by: Waseda, Futa, et al.
Published: (2025)

Directed-Tokens: A Robust Multi-Modality Alignment Approach to Large Language-Vision Models
by: Truong, Thanh-Dat, et al.
Published: (2025)

Source-Free Domain Adaptation with Diffusion-Guided Source Data Generation
by: Chopra, Shivang, et al.
Published: (2024)

Reframing Long-Tailed Learning via Loss Landscape Geometry
by: Chen, Shenghan, et al.
Published: (2026)

Beyond Loss Values: Robust Dynamic Pruning via Loss Trajectory Alignment
by: Qin, Huaiyuan, et al.
Published: (2026)

Proximal Vision Transformer: Enhancing Feature Representation through Two-Stage Manifold Geometry
by: Yun, Haoyu, et al.
Published: (2025)

Revisiting Mixout: An Overlooked Path to Robust Finetuning
by: Aminbeidokhti, Masih, et al.
Published: (2025)

Refining Text-to-Image Generation: Towards Accurate Training-Free Glyph-Enhanced Image Generation
by: Lakhanpal, Sanyam, et al.
Published: (2024)

Vision-Language Feature Alignment for Road Anomaly Segmentation
by: He, Zhuolin, et al.
Published: (2026)

What Makes VLMs Robust? Towards Reconciling Robustness and Accuracy in Vision-Language Models
by: Nie, Sen, et al.
Published: (2026)

Analysing the Robustness of Vision-Language-Models to Common Corruptions
by: Usama, Muhammad, et al.
Published: (2025)

On the Robustness of Medical Vision-Language Models: Are they Truly Generalizable?
by: Imam, Raza, et al.
Published: (2025)

Contextual Self-paced Learning for Weakly Supervised Spatio-Temporal Video Grounding
by: Kumar, Akash, et al.
Published: (2025)

CLAP4CLIP: Continual Learning with Probabilistic Finetuning for Vision-Language Models
by: Jha, Saurav, et al.
Published: (2024)

TRACER: Persistent Regularization for Robust Multimodal Finetuning
by: Asadollahzadeh, Hesam, et al.
Published: (2026)

VLM-RobustBench: A Comprehensive Benchmark for Robustness of Vision-Language Models
by: Saxena, Rohit, et al.
Published: (2026)

On the Adversarial Robustness of 3D Large Vision-Language Models
by: Liu, Chao, et al.
Published: (2026)