The 2024 IEEE Information Theory Workshop (2024 IEEE ITW) was held in Shenzhen from November 24 to 28. It was jointly hosted by the Institute of Data and Information of Tsinghua Shenzhen International Graduate School (Tsinghua SIGS) and IEEE Information Theory Society. IEEE-ITW‘24 is the top international scholar conference on the information theory field and it solicits and welcomes original contributions on the frontiers of information theory, statistics, coding theory, and their applications, as well as interactions between information theory and other fields such as data science, biology, and signal processing.
Shao-Lun Huang, an associate professor at Tsinghua SIGS, Chao Tian, an associate professor at Texas A&M University, and Xiao-Ping Zhang, a Fellow of the Canadian Academy of Engineering, Tsinghua Pengrui Professor, and a professor at Tsinghua SIGS co-chaired the workshop. Sponsored by Huawei Technologies Co., Ltd., Chaspark, and MDPI, the workshop attracted the participation of more than 200 experts, scholars, and students in the field of information theory from more than 20 countries and regions around the world.
The workshop received 201 submissions. After a strict review procedure, 132 papers were accepted, with an acceptance rate of 65.67%. The workshop lasted for five days, including 4 plenary talks, 4 tutorials, 136 academic paper reports, a Professor Ning Cai Memorial Session, a Huawei Industrial Session, and many other activities.
Two tutorials were arranged in the morning and afternoon of November 24, respectively, teaching the converse bound proving techniques for information identification communication, characterizing the generalization error of machine learning algorithms via information measures, mathematical analysis, network algorithms, and applications to epidemics and infodemics, and theory on training dynamics of transformers attracted many participants. The four tutorials extensively discussed cutting-edge issues in information theory research, showing everyone the unique charm of information theory and communication-related technologies in scientific research and application issues.
On the morning of November 25, general co-chairs Shao-Lun Huang and Xiao-Ping Zhang delivered opening speeches for this workshop.
Shao-Lun Huang hosts the opening ceremony and delivers a welcome speech
Xiao-Ping Zhang delivers opening remarks
In the first plenary talk, Professor Aaron Wagner from Cornell University discussed the channel simulation and synthesis problem in depth. His talk showcased an improved scheme that halves the logarithmic term for some channels and eliminates it for all others.
Aaron Wagner gives a plenary talk
A memorial session for Professor Ning Cai was held on the afternoon of November 25, hosted by Masahito Hayashi, Presidential Chair Professor of the Chinese University of Hong Kong (Shenzhen) and Chief Research Scientist of the Shenzhen International Quantum Academy. He also reviewed his scientific collaboration experience with Professor Ning Cai through photo slides and video materials. Qi Cao, an assistant professor at Xi'an University of Electronic Science and Technology, Xuan Guang, a professor at Nankai University, Yanlin Geng, a professor at Xi'an University of Electronic Science and Technology, Raymond Yeung, Choh-Ming Li Professor of Information Engineering, Co-Director, Institute of Network Coding of the Chinese University of Hong Kong also talked about some past events and scientific research cooperation experiences with Professor Ning Cai. Professor Ning Cai's research results have had a profound impact on the fields of network coding and information science. Although he has passed away, his academic results and influence will be passed down forever. The academic legacy he left behind will inspire scholars to continue to pursue knowledge and make greater contributions to the cause of science.
Ning Cai Memorial Session
On the morning of November 26, plenary speaker Prakash Narayan, Professor at University of Maryland, College Park explained the shared information as a potential measure of mutual dependence among multiple jointly distributed discrete random variables in detail. The origins of shared information, its properties, and specific operational meanings were described.
Prakash Narayan gives a plenary talk
On the morning of November 27, Yuejie Chi, Sense of Wonder Group Endowed Professor in AI Systems at Carnegie Mellon University gave a plenary talk, in which she introduced the non-asymptotic theory towards understanding the data generation process of diffusion models and solution of the ill-posed inverse problems motivated by this theory. Along the way, applications in materials science were demonstrated to illustrate the promise of diffusion models in scientific applications.
Yuejie Chi gives a plenary talk
On the morning of November 28, Guangyue Han, Head of Department, Professor at The University of Hong Kong delivered a plenary lecture on a stochastic calculus approach to continuous-time Gaussian channels. In this talk, he explained several challenges in the continuous-time Gaussian channels problem and highlighted several key advantages of the stochastic calculus method in the analysis of this problem.
Guangyue Han gives a plenary talk
While this year’s workshop has called an end, the information theory community, both in China and globally, will continue to delve into academic research and continuously promote technological innovation and innovative development in the information area. We believe that the success of this workshop will also inject new vitality into the innovation and development of Tsinghua SIGS and Shenzhen in related fields.
Written by Tianren Peng, Shengnan Li
Edited by Yuan Yang
Reviewed by Chen Chaoqun
Layout by Yang Jun
