EAR: Erasing Concepts from Unified Autoregressive Models
Open Access
- 9 November 2026
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 1733-1742
Abstract
No abstract availableFunding Information
- National Natural Science Foundation of China (62576179, 62506178)
- Natural Science Foundation of Inner Mongolia Autonomous Region of China (2025JQ012)
This publication has 12 references indexed in Scilit:
- Lumina-mGPT: Flexible Photorealistic Autoregressive Text-to-Image GenerationInternational Journal of Computer Vision, 2026
- Set You Straight: Auto-Steering Denoising Trajectories to Sidestep Unwanted Concepts2025
- STEREO: A Two-Stage Framework for Adversarially Robust Concept Erasing from Text-to-Image Diffusion Models2025
- Multi-View User Preference Modeling for Personalized Text-to-Image GenerationIEEE Transactions on Multimedia, 2025
- Active Supervised Cross-Modal RetrievalIEEE Transactions on Pattern Analysis and Machine Intelligence, 2025
- Defensive Unlearning with Adversarial Training for Robust Concept Erasure in Diffusion Models2024
- Diffusionics2024
- Erasing Concepts from Diffusion Models2023
- AI Art and its Impact on Artists2023
- GenEval: An object-focused framework for evaluating text-to-image alignment2023