Toward Next Generation AI Systems Beyond Lingual Intelligence
Abstract
The success of large language models (LLM) like OpenAI's GPTs has fascinated us, with their astounding capabilities in a wide range of tasks such as genius-level standard test performance, IMO-level math reasoning, encyclopedic-like question-answering, and human-like ability in conversation and writing. However, how such systems are produced and where they go next remain a subject not very accessible to the broader academic and industrial community.
In this talk, I will first share our recent experience on producing industrial-scale LLMs from scratch, introducing a collection of state-of-the-art LLMs we produced in a university team, such as the K2, Jais (Arabic), and vicuna. Then I will discuss limitation of current LLMs in capabilities such as embodied reasoning, physical reasoning, social reasoning, strategic planning, and a pathway toward such capabilities through next generation systems such as World Models capable of instructible and steerable creation (of any next-world resultant from agentic action), and simulative reasoning (thought experiments over alternative next-worlds) in a virtual universe; and Agent Models capable of living, acting, planning, and strategizing in such a universe. My newly sparked attention to philosophy may also land us on some fun debate about what is "AGI" and what is "agency", which may help to invite new ideas from audience.
Biography
Professor Eric Xing is the President of the Mohamed bin Zayed University of Artificial Intelligence, and a Professor of Computer Science at Carnegie Mellon University. His main research interests are the development of machine learning and statistical methodology, and large-scale distributed computational system and architectures, for solving problems involving automated learning, reasoning, and decision-making in in artificial, biological, and social systems. In recent years, he has been focusing on building large language models, world models, agent models, and foundation models for biology.
Prof. Xing has served on the editorial boards of several leading journals including JASA, AOAS, JMLR; was a recipient of several awards including NSF Career, Sloan, Carnegie Science Award, and best papers in conferences such as ACL, ISMB, NeurIPS, and OSDI; and is a fellow of several societies including AAAI, ACM, ASA, IEEE, and IMS.