Jürgen Schmidhuber
Title: RNNaissance
Abstract:
Our deep learning artificial neural networks have won numerous contests in pattern recognition and machine learning. They are now widely used by the world’s most valuable public companies. In particular, Long Short-Term Memory (LSTM) Recurrent Neural Networks (RNNs) are very useful not only for speech recognition but also for Computational Language Learning. I will discuss state-of-the-art results in numerous applications.
Biography:
Since age 15 or so, the main goal of professor Jürgen Schmidhuber (pronounce: You_again Shmidhoobuh) has been to build a self-improving Artificial Intelligence (AI) smarter than himself, then retire. He has pioneered self-improving general problem solvers since 1987, and Deep Learning Neural Networks (NNs) since 1991. The recurrent NNs developed by his research groups at the Swiss AI Lab IDSIA (USI & SUPSI) & TU Munich were the first to win official international contests. They have revolutionized handwriting recognition, speech recognition, machine translation, image captioning, and are now available to over a billion users through Google, Microsoft, IBM, Baidu, and many other companies. DeepMind is heavily influenced by his lab's former students (including 2 of DeepMind's first 4 members and their first PhDs in AI, one of them co-founder, one of them first employee). His team's Deep Learners were the first to win object detection and image segmentation contests, and achieved the world's first superhuman visual classification results, winning nine international competitions in machine learning & pattern recognition (more than any other team). They also were the first to learn control policies directly from high-dimensional sensory input using reinforcement learning. His research group also established the field of mathematically rigorous universal AI and optimal universal problem solvers. His formal theory of creativity & curiosity & fun explains art, science, music, and humor. He also generalized algorithmic information theory and the many-worlds theory of physics, and introduced the concept of Low-Complexity Art, the information age's extreme form of minimal art. Since 2009 he has been member of the European Academy of Sciences and Arts. He has published 333 peer-reviewed papers, earned seven best paper/best video awards, the 2013 Helmholtz Award of the International Neural Networks Society, and the 2016 IEEE Neural Networks Pioneer Award. He is also president of NNAISENSE, which aims at building the first practical general purpose AI.
Fernanda Ferreira
Title: Human Processing of Disfluent Speech: Basic Findings, Theoretical Approaches, and Implications for Natural Language Processing
Abstract:
Disfluencies occur in human speech at the rate of about one per minute; therefore, any adequate theory of human language comprehension must explain how listeners process utterances containing them. Our theoretical approach is based on a 15-year program of research that has uncovered a number of fundamental mechanisms enabling humans to process disfluencies efficiently, including mechanisms that are backward looking (reanalysis of the input) and ones that are anticipatory or forward looking (prediction). This presentation will review the theory, the evidence that supports it, and the outstanding questions that are currently being investigated. I will also consider implications for refining NLP systems, which must be robust to speaker error and which should be capable of adapting to characteristics of particular speakers and language communities.
Biography:
Fernanda Ferreira is Professor of Psychology and Member of the Graduate Group in Linguistics at the University of California, Davis. She obtained her Ph.D. in Cognitive Psychology in 1988 from the University of Massachusetts, Amherst, and prior to moving to UC Davis in 2015, she held faculty positions at Michigan State University and the University of Edinburgh. She has published over 100 papers and her research has been funded by the NSF and the NIH in the US, and the ESRC in the UK. She served as Editor in Chief of the Journal of Experimental Psychology: General, and she is currently an Associate Editor of Cognitive Psychology and of Collabra, an Open Access journal recently launched by University of California Press. She is a Fellow of the American Psychological Society and the Royal Society of Edinburgh, and she is currently an elected member of the Psychonomic Society's Governing Board.