Understanding the Training and Inference of Reinforcement Learning

大模型技术在快手AI的应用与落地

Data Attribution: A Principled Data-Centric Approach for Trustworthy AI

深度学习理论分析：从数据和算法的角度出发

The Price Performance of Performance Models

AI's Certain Uncertainty Problem

From Simulated Subjectivity to Collective Consciousness in Large Language Models

Overview of frontier AI safety issues and work from the UK

Abstract:Frontier Al systems that are approaching AGl could pose significant risksto society. This talk will provide an overview of the work done by the Ukgovernment's Al safety institute on investigating potential risks posecby these Al systems, and the work they are doing to address them.