学术报告会 | Understanding the World with AI: Training and Validating AI Systems Using Synthetic Data

  • 学术报告会 | Understanding the World with AI: Training and Validating AI Systems Using Synthetic Data已关闭评论
  • 2,460

时间

2019年8月12日14:00

地点

上海交通大学闵行校区软件大楼5楼 人工智能研究院 500会议室

摘要

The talk with start with an overview of the German Research Center for Artificial Intelligence (DFKI), the largest AI research center worldwide and some of our current activities across AI, both on the technical and on the political side. In the second part I will then discus in more depth one of our key research areas -- Digital Reality: Despite the fact that the world around us is highly complex, AI systems must be able to reliably make accurate decisions that in many cases could even affect human lives (e.g. Autonomous Driving). We propose an approach that does not only rely on real data but that learns models of the real world and then uses synthetic sensor data generated via simulations for the training and -- even more importantly -- the validation of AI Systems. This is extended by a continuous process of validating the models against the real world in order to extend and adapt them to a changing environment. A key element of this approach is the accurate simulation not only of the environmental model but also of any relevant sensor used by an AI system. Besides cameras, lidar, and others, radar becomes increasingly important in the context of autonomous driving. Finally, this approach requires to develop simulation techniques that provide some guarantees about the accuracy of the simulation.

个人简介

Philipp Slusallek is Scientific Director at the German Research Center for Artificial Intelligence (DFKI), where he heads the research area on Agents and Simulated Reality. At Saarland University he has been a professor for Computer Graphics since 1999, a principle investigator at the German Excellence-Cluster on “Multimodal Computing and Interaction” since 2007, and was Director for Research at the Intel Visual Computing Institute 2009-2017. Before coming to Saarland University, he was a Visiting Assistant Professor at Stanford University. He is a fellow of Eurographics, a member of acatech (German National Academy of Science and Engineering), and a member of the European High-Level Expert Group on Artificial Intelligence. HProf. Slusallek co-founded the European AI Initiative CLAIRE (Confederation of Laboratories for Artificial Intelligence Research in Europe, claire-ai.org) in 2018. He originally studied Physics in Frankfurt and Tübingen (Diploma/M.Sc.) and got his PhD in Computer Science from Erlangen University. His research covers a wide range of topics including artificial intelligence in a broad sense, simulated/digital reality, real-time realistic graphics, high-performance computing, motion synthesis, novel programming models for CPU/GPU/FPGA, computational science, and others.