Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore a keynote presentation from the Future Technologies Conference 2024 where Professor Peter Triantafillou, a distinguished computer scientist from the University of Warwick, delves into the emerging field of machine unlearning. Learn about the critical challenges facing AI systems trained on problematic datasets containing biased, obsolete, or sensitive information, and discover how these issues impact crucial societal systems in healthcare, justice, and energy infrastructure. Understand the innovative process of "unlearning" - a resource-efficient approach to removing harmful effects from AI models without complete retraining. Gain insights into the complexities of defining and measuring unlearning across various AI applications, including image classification, natural language processing, and large language models. Master the concepts behind efficient unlearning algorithms that balance model performance with privacy concerns, accuracy requirements, and generalization capabilities. Examine how this cutting-edge research addresses AI safety and critical infrastructure protection through practical examples and expert analysis from a renowned researcher whose work has garnered multiple best paper awards and who has served in leadership roles across prestigious computer science conferences.
