Understanding and Mitigating Copying in Diffusion Models
Offered By: Google TechTalks via YouTube
Course Description
Overview
Explore the challenges and solutions surrounding content replication in diffusion models through this Google TechTalk presented by Gowthami Somepalli. Delve into various frameworks for detecting replication and learn about the study that identified memorization in Stable Diffusion 1.4. Examine the factors contributing to memorization in diffusion models, including the often-overlooked role of text conditioning. Discover proposed techniques for reducing data replication during both training and inference times. Gain valuable insights into the complexities of maintaining originality in AI-generated art and graphic design as cutting-edge diffusion models continue to produce high-quality, customizable images for commercial purposes.
Syllabus
Understanding and Mitigating Copying in Diffusion Models
Taught by
Google TechTalks
Related Courses
Introduction to Artificial IntelligenceStanford University via Udacity Computer Vision: The Fundamentals
University of California, Berkeley via Coursera Computational Photography
Georgia Institute of Technology via Coursera Digital Signal Processing
École Polytechnique Fédérale de Lausanne via Coursera Creative, Serious and Playful Science of Android Apps
University of Illinois at Urbana-Champaign via Coursera