paper-with-me

홈 › Papers

Bridging Mode Connectivity in Loss Landscapes and Adversarial Robustness

2020-04-30 · ICLR 2020 1 · Pu Zhao, Pin-Yu Chen, Payel Das, Karthikeyan Natesan Ramamurthy, Xue Lin

Mode connectivity provides novel geometric insights on analyzing loss landscapes and enables building high-accuracy pathways between well-trained neural networks. In this work, we propose to employ mode connectivity in loss landscapes to study the adversarial robustness of deep neural networks, and provide novel methods for improving this robustness. Our experiments cover various types of adversarial attacks applied to different network architectures and datasets. When network models are tampered with backdoor or error-injection attacks, our results demonstrate that the path connection learned using limited amount of bonafide data can effectively mitigate adversarial effects while maintaining the original accuracy on clean data. Therefore, mode connectivity provides users with the power to repair backdoored or error-injected models. We also use mode connectivity to investigate the loss landscapes of regular and robust models against evasion attacks. Experiments show that there exists a barrier in adversarial robustness loss on the path connecting regular and adversarially-trained models. A high correlation is observed between the adversarial robustness loss and the largest eigenvalue of the input Hessian matrix, for which theoretical justifications are provided. Our results suggest that mode connectivity offers a holistic tool and practical means for evaluating and improving adversarial robustness.

📄 PDF Abstract BibTeX arXiv:2005.00060

Code (3)

IBM/model-sanitization 공식 구현 pytorch
BililiCode/ARGD pytorch
bboylyg/NAD pytorch

Tasks

Adversarial Robustness

Methods 이 논문이 사용한 방법론

Repair 설명 없음

Similar Papers 제목 키워드 기반

Optimizing Mode Connectivity via Neuron Alignment

2020-09-05 · NeurIPS 2020 12 · N. Joseph Tatro, Pin-Yu Chen, Payel Das, Igor Melnyk 외

The loss landscapes of deep neural networks are not well understood due to their high nonconvexity. Empirically, the local minima of these loss functions can be connected by a learned curve in model space, along which th…

Black holes and the loss landscape in machine learning

2023-06-26 · Pranav Kumar, Taniya Mandal, Swapnamay Mondal

Understanding the loss landscape is an important problem in machine learning. One key feature of the loss function, common to many neural network architectures, is the presence of exponentially many low lying local minim…

Mode Connectivity Beyond Classifiers: Evidence from Generative and Contrastive Models

2026-08-31 · Chengzheyi Yao, Yongzhao Zhang, Yongding Tian arxiv

The loss landscape of Deep Neural Networks (DNNs) exhibits highly complex and non-convex properties. Recent studies have revealed the phenomenon of mode connectivity, demonstrating that independently trained network mode…

Understanding Adversarial Robustness Through Loss Landscape Geometries

2019-07-22 · Vinay Uday Prabhu, Dian Ang Yap, Joyce Xu, John Whaley

The pursuit of explaining and improving generalization in deep learning has elicited efforts both in regularization techniques as well as visualization techniques of the loss surface geometry. The latter is related to th…

Adversarial RobustnessData Augmentation

Phase transitions reveal hierarchical structure in deep neural networks

2025-12-05 · Ibrahim Talha Ersoy, Andrés Fernando Cardozo Licha, Karoline Wiesner arxiv

Training Deep Neural Networks relies on the model converging on a high-dimensional, non-convex loss landscape toward a good minimum. Yet, much of the phenomenology of training remains ill understood. We focus on three se…