TopicsAlphaGO, Model-Based Reinforcement Learning
Ioannis Antonoglou’s papers:
Constructing agents with planning capabilities has long been one of the main challenges in the pursuit of artificial intelligence. Tree-based planning methods have enjoyed huge success in challenging domains, such as chess and Go, where a perfect simulator is available. However, in real-world problems the dynamics governing the environment are often complex and unknown. In this work we present the MuZero algorithm which, by combining a tree-based search with a learned model, achieves superhuman performance in a range of challenging and visually complex domains, without any knowledge of their underlying dynamics. MuZero learns a model that, when applied iteratively, predicts the quantities most directly relevant to planning: the reward, the action-selection policy, and the value function. When evaluated on 57 different Atari games – the canonical video game environment for testing AI techniques, in which model-based planning approaches have historically struggled – our new algorithm achieved a new state of the art. When evaluated on Go, chess and shogi, without any knowledge of the game rules, MuZero matched the superhuman performance of the AlphaZero algorithm that was supplied with the game rules.
TopicsTensorFlow 2.0, Data Analysis, Machine Learning
Paige Bailey, Mountain View, CA, USA
TensorFlow Developer Advocate, Google
Paige Bailey is a TensorFlow Developer Advocate at Google, based in Mountain View, CA. Prior to joining Google, Paige worked as a senior software engineer in the office of the Azure CTO; as a Cloud Developer Advocate for machine learning at Microsoft; and as a data scientist for Chevron in Houston, TX.
Paige has over a decade of experience using Python for data analysis, five years of experience doing machine learning – and can’t wait to show you about the new capabilities in TensorFlow 2.0.
TopicsComputer Vision, Deep Learning.
Professor Cipolla is Professor of Information Engineering at the University of Cambridge (since 2000) and Director of Toshiba’s Cambridge Research Laboratory (since 2007).
After reading Engineering at the University of Cambridge (Queens’ College, 1984) he completed his graduate research in Computer Vision and Robotics at the University of Oxford (Balliol College, 1991) where he received D.Phil. degree for his work on 3D reconstruction from smooth 2D contours.
Cipolla’s research interests are in the reconstruction, registration and recognition of three-dimensional objects from images. These include novel algorithms for the recovery of accurate 3D shape, visual localisation and semantic segmentation. Computer vision technology from his group is being exploited in new products by Toshiba (face recognition for access control in varying illumination and a gesture interface for laptops) and Wayve Technologies (semantic segmentation for autonomous driving).
He has authored two books: Active Visual Inference of Surface Shape in 1995 and Visual Motion of Curves and Surfaces (with Peter Giblin) in 2000; edited twelve books on computer vision and published over 400 articles in computer vision and related fields. See https://mi.eng.cam.ac.uk/~cipolla/publications_selected.htm
Cipolla was elected a Fellow of the Royal Academy of Engineering in 2010; a Distinguished Fellow of the British Machine Vision Association in 2013; and a Fellow of the International Association for Pattern Recognition in 2020.
The last decade has seen a revolution in the theory and application of computer vision and machine learning. I will begin with a brief review of some of the fundamentals with a few examples of the reconstruction, registration and recognition of three-dimensional objects and their translation into novel commercial applications. I will then introduce some recent results from real-time deep learning systems that exploit geometry and compute model uncertainty. Understanding what a model does not know is a critical part of safe machine learning systems. New tools, such as Bayesian deep learning, provide a framework for understanding uncertainty in deep learning models, aiding interpretability and safety of such systems. Additionally, knowledge of geometry is an important consideration in designing effective algorithms. In particular, we will explore the use of geometry to help design networks that can be trained with unlabelled data for human body pose and shape recovery.
TopicsOptimization, Complex Networks & Data Science
Panos M. Pardalos serves as distinguished professor of industrial and systems engineering at the University of Florida. Additionally, he is the Paul and Heidi Brown Preeminent Professor of industrial and systems engineering. He is also an affiliated faculty member of the computer and information science Department, the Hellenic Studies Center, and the biomedical engineering program. He is also the director of the Center for Applied Optimization. Pardalos is a world leading expert in global and combinatorial optimization. His recent research interests include network design problems, optimization in telecommunications, e-commerce, data mining, biomedical applications, and massive computing.
Abstract: Artificial Intelligence (AI) has been a fundamental component of many activities in biomedicine in recent years. In this lecture we first summarize some of the major impacts of AI tools in biomedicine and discuss future developments and limitations. In the second part of the talk we present details on our ongoing research on brain dynamics.
TopicsNatural Language Processing, Conversational AI, Spoken Dialogue Systems, Dialog, Natural Language Generation
Verena Rieser is a professor at Heriot-Watt University where she leads research on Natural Language Processing with applications in Natural Language Generation and Spoken Dialogue Systems. She is also a co-founder of the Conversational AI company Alana AI, a Leverhulme Trust Senior Research Fellow awarded by the Royal Society, and PI of several funded research and industry awards.