Workshop is organized by Viraj Prabu, Prithvijit Chattopadhyay, sriram Yenamandra, Krish Kabra, Hao Liang, Guha Balakrishnan, Pietro Perona
Posts by Judy Hoffman
Workshop Focus: Traditional ML auditing relies on manually-labeled data with limitations. Our workshop explores how generative techniques can create synthetic data that allows for controlled attribute manipulation to analyze performance, bias, and failure modes in ML systems. ๐
๐ Excited about how generative AI can power experimental (not just observational) audits of ML systems that reveal actionable insights into performance and bias?
Join us: Experimental Model Auditing with Controllable Synthesis workshop @cvprconference.bsky.social!
sites.google.com/view/emacs20...
๐ฎโ๐จ
๐จ NeurIPS 2024 ๐จHow robust are AI-Generated Image Detectors?
๐ค Can they detect various magnitudes of image augmentations?
๐ก Does performance fluctuate across scenes?
๐ Find out with Semi-Truths: 1.5 million images for the targeted evaluation of AI-generated images. arxiv.org/abs/2411.07472
Hi! Who's here from the computer vision or AI communities?