主 講 人:Bj?rn W. Schuller教授
地 點:物信學院祥聯廳
主 辦 方:物理與信息工程學院
開始時間:2018-05-14 09:30
Bj?rn W. Schuller教授簡介:
IEEE FELLOW,博士畢業于德國慕尼黑工業大學,德國奧格斯堡大學W3終身教授(德國最高教職),現任英國帝國理工學院Reader。長期致力于人工智能、音頻識別、情感計算、機器學習的相關算法和應用研究。曾在國際著名學術期刊上發表論文100余篇,在國際知名學術會議上發表論文400余篇。并成功舉辦了十屆INTERSPEECH Computational Paralinguistics Challenge。擔任大會主席:IEEE ACII 2019, IEEE ACII Asia 2018, ACM ICMI 2014
報告內容簡介:
Artificial Intelligence has recently gained tremendously both in performance and popularity owing to a large extent to the advances in Deep Learning and exploitation of increasingly Big Data. This has opened up new opportunities in the domain of human-centred A.I. The idea of the latter is to enable next-generation intelligent systems to considerably better understand humans, such as by the inclusion of socioemotional artificial intelligence. In this respect, recent advances in modelling and profiling of users by deep learning will be presented including benchmarks and findings from several challenge events organised by the presenter at leading outlets in the field. At the same time, ensuring better understanding of A.I. by humans is crucial for the acceptance and optimal usage and exploitation of A.I. by humans. Along these lines, increasde of transparency and “white-boxing“ in deep learning frameworks will be discussed. Finally, deep learning empowered computational creativity will be touched upon as third component of human-centered A.I. before showing opportunities in terms of application domains. The talk will be concluded by a presentation of current challenges in the field.