Demystifying AI interpretability | VB Transform 2020
3Demystifying AI interpretability; Improving accuracy and predictability of AI models using reinforcement learning
Reinforcement Learning (RL) is a machine learning technique that is being quickly adopted by many organizations to solve large and complex problems, where labeled datasets are not readily available. Since RL learns by a continuous process of rewards and punishments on the actions it takes, it is able to train algorithms that interact with new environments. RL can also help in designing more explainable and accurate AI, which is one of the most important considerations in designing AI systems today. In this session, Ion Stoica, the director of RISELab (UC Berkeley) will discuss some of the most relevant advances and applications in this space.
Ion Stoica, Co-founder, Rise Labs (UC Berkeley),
In conversation with Jed Dougherty, VP of Field Engineering, Dataiku