Big Models, Big Impact? Machine Learning in the Age of Large Language and Foundation Models

Info

Haizhou Li

LLM Innovation with Chinese Characteristics

This talk examines the strategic and technical path China’s large language model (LLM) industry is forging under unique constraints, including chip sanctions and blockade of access by closed source LLMs. It is believed that these constraints have catalyzed a distinctive innovation model, shifting the focus from brute-force scaling and to an approach centered on efficiency, open-source, and application.

We will also discuss the public acceptance of AI in China. China’s LLM strategy offers a compelling alternative innovation paradigm where strategic constraints can foster creativity.

Haizhou Li is a U Bremen Excellence Chair Professor. He is also with National University of Singapore and The Chinese University of Hong Kong, Shenzhen. Professor Li’s research interests include speech information processing, natural language processing, and neuromorphic computing. He served as the Editor-in-Chief of IEEE/ACM TRANSACTIONS ON AUDIO, SPEECH AND LANGUAGE PROCESSING (2015-2018), the President of the international Speech Communication Association (2015-2017), the President of Asia Pacific Signal and information Processing Association (2015-2016), the President of the Asian Federation of Natural Language Processing (2017-2018), the Vice President of IEEE Signal Processing Society (2024-2026). He was the General Chair of ACL 2012, INTERSPEECH 2014, IEEE ICASSP 2022, and APSIPA Annual Summit and Conference 2025. Professor Li was the recipient of National Infocomm Awards 2002, and the President’s Technology Award 2013 in Singapore. He was named Nokia Professor in 2009, IEEE Fellow in 2014, ISCA Fellow in 2018, and Fellow of the Academy of Engineering Singapore in 2022. 

Jürgen Schmidhuber

The Neural World Model Boom

The concept of a mental model of the world—a world model—dates back millennia. Plato suggested that we recognize objects by recollecting internal blueprints or templates, today often called internal representations. Aristotle wrote that phantasia or mental images allow humans to imagine the future and to plan action sequences by mentally manipulating images in the absence of the actual objects. Only 2,370 years later—a mere blink of an eye by cosmical standards—we are witnessing a boom in world models based on deep artificial neural networks for artificial intelligence in the physical world. New startups on this are emerging. To explain what's going on, I'll take you on a little journey through the history of general purpose neural world models since 1990. Then we'll take a look at the future of this field. 

Jürgen Schmidhuber is the Director of the KAUST AI Initiative and Co-Chair of the Center of Excellence for Generative AI at King Abdullah University of Science and Technology (KAUST). He is also Scientific Director of the Swiss AI Lab, IDSIA, and was Professor of Artificial Intelligence at the University of Lugano (USI) from 2009 to 2021. Professor Schmidhuber's research interests include deep learning, artificial neural networks, recurrent neural networks and Long Short-Term Memory, artificial curiosity, and neuromorphic computing. He is Co-Founder and former President of NNAISENSE and a board member of Delvitech. He was the recipient of the Helmholtz Award of the International Neural Network Society (2013), the IEEE CIS Neural Networks Pioneer Award (2016), and the Gödel Prize, among other honors. He is a member of the European Academy of Sciences and Arts, and was named one of the "Top 100 Leaders of Switzerland" by BILANZ magazine in 2018. He earned his Diploma in Computer Science and Mathematics from the Technical University of Munich (TUM), and his Ph.D. in Computer Science from TUM in 1991, followed by his Habilitation there in 1993. He is often described in the media as one of the pioneers of modern AI, having laid foundational groundwork for what is now known as Generative Adversarial Networks, unnormalized linear Transformers, and self-supervised pre-training.

Zur Anmeldung »