Abstract

Metaverse is a term used to describe a virtual world where humans live, learn, work, and interact with each other through digital avatars. Many envision the metaverse as the next evolution of the Internet: the future of social interaction, entertainment and commerce where humans experience and see life in ways they could not in the physical world. Recent advances in virtual reality, augmented reality and mixed reality technology are among key technology that provide truly immersive experiences in 3D virtual space. Despite its potential, the metaverse is still at early stage and in the foreseeable future no single platform will dominate every sectors. This tutorial presents how a virtual university campus can be built by academic sector and what benefits can the metaverse bring to academic world. An example of a virtual university campus, called intaniaverse.com, will be used as a practical study case. This tutorial will be useful for those who seek to construct a metaverse by themselves.

Abstract

Today, in order to realize comfortable life and economic development, optimization processing using a large amount of complex data is promising. Optimization processing with large amount of data using conventional computing technologies require a large amount of processing time. An annealing machine, a new-paradigm computing technology inspired by a quantum computer, was proposed to accelerate the optimization processing. Now, many implementations of annealing machines are developed and commercialized in various fields.

In this talk, the necessity of annealing machines is discussed for an actual social implementation and examples of real applications are also introduced. Then CMOS annealing, which is one implementation of annealing machine using CMOS process, is introduced. The annealing machine is newly developed technology and there are many challenges to spread the technology. The challenges are also explained.

Abstract

Self-driving vehicle perception and control have made dramatic progress in the last several years, and many auto vendors have pledged large-scale commercialization in a 2-3 year time frame. These controllers use a variety of approaches but recent successes suggest that neural networks will be widely used in self-driving vehicles. Classical AI systems involve carefully-crafted features and representations, while one of the new powers of deep learning methods is the ability to learn very effective latent representations from data. There have been a variety of approaches, which depend on the modular perception-prediction-planning-control pipeline, where each module can be built using deep learning methods. In this talk, I will introduce some of Waymo Research's recent work on deep learning for autonomous driving at scale and the variety of challenges towards fully self-driving ride on public roads.