Computer Vision

Mining Attribute Subspaces for Efficient Fine-tuning of 3D Foundation Models
With the emergence of 3D foundation models, there is growing interest in fine-tuning them for downstream tasks, where LoRA is the …
Yu Jiang, Hanwen Jiang, Ahmed Abdelkader, Wen-Sheng Chu, Brandon Y. Feng, Zhangyang Wang, Qixing Huang
Cite Project arXiv
Mull-Tokens: Modality-Agnostic Latent Thinking
Reasoning goes beyond language; the real world requires reasoning about space, time, affordances, and much more that words alone cannot …
Arijit Ray, Ahmed Abdelkader, Chengzhi Mao, Bryan A. Plummer, Kate Saenko, Ranjay Krishna, Leonidas Guibas, Wen-Sheng Chu
Cite Project Webpage arXiv Code
Mull-Tokens: Modality-Agnostic Latent Thinking
Marginalized Bundle Adjustment: Multi-View Camera Pose from Monocular Depth Estimates
Structure-from-Motion (SfM) is a fundamental 3D vision task for recovering camera parameters and scene geometry from multi-view images. …
Shengjie Zhu, Ahmed Abdelkader, Mark J. Matthews, Xiaoming Liu, Wen-Sheng Chu
Cite Project Webpage OpenReview arXiv Code
Computer Vision
Perception and Reasoning from Pixels
Detection as Regression: Certified Object Detection by Median Smoothing
Despite the vulnerability of object detectors to adversarial attacks, very few defenses are known to date. While adversarial training …
Ping-yeh Chiang, Michael Curry, Ahmed Abdelkader, Aounon Kumar, John Dickerson, Tom Goldstein
Cite Project NeurIPS arXiv Code
Certified Defenses for Adversarial Patches
Adversarial patch attacks are among one of the most practical threat models against real-world computer vision systems. This paper …
Ping-yeh Chiang, Renkun Ni, Ahmed Abdelkader, Chen Zhu, Chris Studor, Tom Goldstein
Cite Project OpenReview arXiv Code
Machine Learning
Mitigating the vulnerabilities of modern machine learning workflows