Tutorial

Generative AI: Techniques and Applications with Large Language Models

Tutorial Time: 13:00-14:30 p.m. PHT (UTC+8), November 5, 2024

Ao Guo, Ph.D. in Computer Science

Graduate School of Informatics, Nagoya University, Japan

Abstract: Generative AI (GAI) has seen rapid advancements and gained immense popularity in recent years. With its extensive model parameters and the ability to learn from vast datasets, GAI is capable of creating new content, such as text, images, music, and videos, which were once considered exclusive to human creativity. Among these, Large Language Models (LLMs) have significantly advanced the capabilities of understanding and generating human-like text, making them a pivotal component of GAI. This tutorial is designed to provide a comprehensive overview of GAI in terms of key techniques and popular applications, with a particular focus on LLMs. We will start with an overview of GAI, highlighting significant milestones and innovations. Key techniques such as transformer architectures, self-attention mechanisms, and Generative Adversarial Networks (GANs) will be explained in detail. We will explore the efficient training methods that enable the scaling and optimization of these models. Additionally, we will present popular applications of GAI across various domains. Finally, we will address challenges and considerations in the field of generative AI with LLMs.

Biography: Ao Guo received his M.S. degree from Huazhong University of Science and Technology, China, in 2016, and his Ph.D. from Hosei University, Japan, in 2021. Since April 2021, he has been a postdoctoral researcher at the Graduate School of Informatics, Nagoya University, Japan. During his Ph.D. research, he proposed an integrative Personal Character Model (PCM) to comprehensively characterize individuals and developed a systematic framework for PCM modeling. From April 2021 to March 2023, Dr. Guo focused on the Interdisciplinary Area project "Studies on Intelligent Systems for Dialogue toward the Human-machine Symbiotic Society". Since April 2024, he has been involved in the Moonshot Research and Development Program project "The Realization of an Avatar-Symbiotic Society where Everyone can Perform Active Roles without Constraint". His current research interests include Persona Character Modeling, Personalized Generative AI, Task-oriented Dialogue Systems, and Autonomous Mobile Conversational Robots. Dr. Guo has received the Best Paper Award at IEEE Cyber, Physical and Social Computing (CPSCom) 2016, the Outstanding Ph.D. Dissertation Award from the IEEE Technical Committee on Hyper-Intelligence (HITC) in 2022, and the Outstanding Paper Award at IEEE CPSCom 2023. Since 2021, he has been serving as TC Secretary for the IEEE HITC.