On July 4, the third session of the Longbin Distinguished Seminar was held at the University of Michigan-Shanghai Jiao Tong University Joint Institute (UM-SJTU JI, JI hereafter). Professor David Atienza from École Polytechnique Fédérale de Lausanne (EPFL) delivered a keynote on “Artificial Intelligence as a Game Changer for Sustainable Computing,” attracting faculty, students, and guests from across SJTU.
JI Associate Dean for Academic Affairs David Hung opened the seminar by welcoming Professor Atienza and highlighting the Longbin Distinguished Seminar as a platform for interdisciplinary exchange and international collaboration. Tenured Associate Professor Weikang Qian introduced Professor Atienza’s achievements in energy-efficient computing and sustainable system design.
In his talk, Professor Atienza explored the challenges of supporting AI workloads—particularly machine learning (ML) and generative AI—in modern data centers. He emphasized that while integrating multiple AI/ML services improves resource use, it also creates intense competition for limited resources and risks overprovisioning. To address this, his team developed machine learning-based solutions that predict cloud workload performance and optimize runtime resource allocation. These techniques enable dynamic voltage and frequency scaling (DVFS) to maximize energy efficiency without compromising performance.
Professor Atienza also presented results from EPFL’s CompuSapien and HeatingBits projects, demonstrating how next-generation data centers can combine deep learning workload analysis, innovative 3D MPSoC architectures, and renewable energy to reuse waste heat at scale. He underscored the importance of modeling sustainability with measurable indicators such as energy footprint, carbon emissions, and cost to achieve balanced design solutions.
As a leading researcher in embedded systems and sustainable computing, Professor Atienza has authored over 400 publications, holds 14 patents, and serves as Associate Vice President at EPFL. He is a Fellow of IEEE and ACM and has received multiple international awards recognizing his contributions to the field.
The seminar concluded with a lively discussion where participants engaged on topics including AI resource management, the robustness of ML models, and the industrialization of waste heat reuse.
David Atienza receives a certificate of appreciation and a commemorative photo of the event from David Hung.
Group photo of event attendees
David Hung noted in his closing remarks that the Longbin Distinguished Seminar continues to bring together global expertise to inspire interdisciplinary innovation and international academic exchange.
