Plenary Talk

By looking back at the technology sector, common phrases such as the Internet of Things, Industry 4.0, Big Data, and Artificial Intelligence dominated the discussions through various fields of research, industry, and economy. Due to the rapid development of parallel processing hardware technology, algorithms, and the corresponding software solutions since 2015, this trend experienced unprecedented acceleration, with the release of the first ChatGPT version in late 2022, highlighting the age of AI. Despite the ongoing development of Artificial Intelligence, the excessive expectations in the technology followed by depression and disappointment as one part of a typical hype cycle, this technology offers a wide range of possibilities for researchers and engineers in the field of theoretical, experimental, and computational vibroacoustics if applied correctly and with care. This presentation examines current developments in artificial intelligence together with risks and opportunities when utilizing the technology to solve problems in the field of vibroacoustic. The importance of high-quality data and its associated nature are stressed, leading to a knowledge-and-experience-enhanced artificial intelligence (keeAI) incorporating problem-specific domain expertise of the developer and engineer, respectively. The presented examples serve as a possible template for future developments and help to facilitate the application of Big Data approaches and Artificial Intelligence beyond the disappointment of excessive expectations.

Doctor Maeder is a fully employed Research Associate at the chair of Vibroacoustics of Vehicles and Machines at the TUM School of Engineering and Design at the Technical University of Munich (TUM), Germany, from which he also received his doctoral degree in Engineering Mechanics with his work “Sound and vibration in a mixed frame - Applications in aeroacoustics and rotor dynamics”. Doctor Maeder is the author and co-author of numerous peer-reviewed journal articles, a book chapter, a patent, and over 50 conference contributions on various research topics in numerical and experimental vibroacoustics. Specifically, this includes applications in the automotive and aerospace sector, civil engineering, medical equipment, and material property identification of monolithic and carbon composite structures. His research is based on the pillars of fundamental theory, numerical methods, experiments, and data-driven approaches to understand wave propagation in solids and fluids and their interaction. In addition to his research, doctor Maeder is teaching courses in “Machine Learning” and “Experimental Vibroacoustics” at TUM and was a guest lecturer for “Noise, Vibration, and Harshness” at Tongji University, Shanghai, China. Besides his commitment as a reviewer for various international journals, Marcus Maeder is an Associated Editor for Numerical and Computational Acoustics for Acta Acustica and the current chair of the Technical Committee on Computational Acoustics within the European Acoustics Association.

In this talk, I will present our recent advances in applying machine learning to underwater acoustic source localization and propagation modeling. Our studies on source localization cover a range of environments, from shallow waters to deep sea scenarios, utilizing both real and synthetic data sets for training. The results demonstrate the superiority of machine learning methods over traditional approaches in handling the environmental uncertainties. We will also discuss some conclusions and the challenging problems encountered in source localization applications. The second focus of this presentation is on acoustic propagation modeling using neural operators. Unlike physics-informed neural network (PINN) methods, neural operators derive underlying relationships primarily from extensive, well-prepared data sets. Instead of learning mappings between finite-dimensional Euclidean spaces, these data-driven neural operators learn mappings between infinite-dimensional function spaces, which is particularly attractive for sound propagation modeling tasks. We will examine the generalization capabilities of neural operators when applied to sound propagation modeling in range-independent shallow water environments.

Dr. Haiqiang Niu is a full Professor at the State Key Laboratory of Acoustics, Institute of Acoustics, Chinese Academy of Sciences. He serves as an associate editor for the Journal of the Acoustical Society of America (JASA) and as a guest editor for the Journal of Marine Science and Engineering (JMSE). He is also a member of the Young Scientist Committee for several journals, including Chinese Physics Letters, Chinese Physics B, Acta Physica Sinica, Physics, and Acta Acustica. Additionally, he is a member of the Youth Innovation Promotion Association of the Chinese Academy of Sciences. Dr. Niu received his Ph.D. in Acoustics from the Institute of Acoustics, Chinese Academy of Sciences, in 2014. From 2015 to 2017, he worked as a postdoctoral researcher at the Scripps Institution of Oceanography, University of California San Diego. He became an associate professor in 2018 and was promoted to full professor in 2021. His research interests include machine learning in ocean acoustics, sparse Bayesian learning in acoustical signal processing, geoacoustic inversion, and ocean acoustical tomography. One of his papers on machine learning has garnered over 200 citations. Dr. Niu has been invited to present at several international conferences.

In this talk, recent advances in acoustic and elastic meta-structures to control noise and vibration are presented aiming at practical applications to real-world problems in our daily life and various industries. The first example is an ultra-light (20 times lighter than existing materials) soundproofing meta-panel to insulate broadband noise generated from electric vehicles, with the aid of negative mass density in low-frequency range (road noise) and negative bulk modulus in high-frequency range (motor noise). The second one is an acoustic meta-liner to insulate noise in a duct while allowing flow, by designing ultra-thin acoustic meta-surface with the consideration of visco-thermal losses in deep-subwavelength-scale Helmholtz resonators. The third one is a meta-damper to suppress vibration in beams or plates using waveguide absorbers based on Archimedean spiral acoustic black holes. For the three meta-structures, we present how to design them theoretically and validate their performance experimentally with a couple of audiovisual demonstrations.

Wonju Jeon is an Associate Professor in the Department of Mechanical Engineering at Korea Advanced Institute of Science and Technology (KAIST) and leading the Wave Energy Control Laboratory. Before joining KAIST in 2014, he worked for National Institute for Mathematical Sciences to bridge the gap between applied mathematics and engineering sciences for 7 years. He received his Ph.D. from KAIST in 2006, with a focus on mathematical theory of acoustic diffraction aiming at reducing fan noise. He aims to develop new and innovative solutions for long-standing problems of sound and vibration occurring in various industries such as home appliances and future mobilities. To overcome the limitation of existing materials or technologies, he proposed new concepts of meta-structures (‘meta’ in Greek means ‘beyond’ in English) such as ‘spiral acoustic black holes’, ‘complex-valued impedance tiles’, ‘ultra-light soundproofing meta-panels’ and ‘sound-absorbing metasurface'. He published papers in the field of sound and vibration, pure and applied mathematics, theoretical biology and applied physics. He gave keynote lectures at The 49th INTER-NOISE (2020) and The 23rd Workshop of the Aeroacoustics Specialists Committee of the Council of European Aerospace Societies (2019). Currently, he serves as Secretary General of The 31st International Congress on Sound and Vibration (2025) and Technical Program Committee of Phononics 2025.