SATELLITE WORKSHOP
ENERGY-EFFICIENT AI ARCHITECTURES WORKSHOP
Colocated with the 2026 IEEE Midwest Symposium on Circuits and Systems (MWSCAS)
OVERVIEW
The goal of the proposed workshop is to bring together experts to discuss recent advances in energy-efficient artificial intelligence (AI). AI market is expected to grow to $5T by 2033, and is expected to create 12-78 million new jobs, as per Anthropic’s 2026 AI index. There has been tremendous growth in AI over last ten years. The number of parameters in current AI models is in the range of 2-6 T, and the cost of energy consumption to train these models is in the range of $500M-$1B+.
Reducing energy consumption in AI agents and data centers is one of the challenges that needs to be addressed by Circuits and Systems researchers. How can we reduce the amount of data needed to train these agents? How can we reduce energy consumption of post-training and inference-time training? What are the approaches to reducing energy consumption in design of low-rank approximation modules? The workshop speakers will discuss solutions to these problems and will identify future directions.
The workshop is free to all participants including the attendees of the 2026 IEEE MWSCAS, but registration is required.
WORKSHOP AGENDA
REGISTRATION
MORNING SESSIONS
-
Domain-specific Energy-efficient Artificial Intelligence AgentsKeshab K. Parhi, University of Minnesota Twin Cities
Bottom-up Domain-Specific SuperintelligenceNiraj K. Jha, Princeton University
Analog In-Memory Computing is a Key Driver of Energy Efficient AINaveen Verma, Princeton University and EnCharge AI
AFTERNOON SESSIONS
-
Physical AI at the Edge: Connecting Perception to Action for On-Device Autonomous IntelligenceTinoosh Mohsenin, Johns Hopkins University
Beyond Efficient Accelerators: System Design Challenges for Sustainable AI
Krishnan Kailas, Independent Researcher and Entrepreneur
How Memory Powers Intelligence: From Cognition to AI Computing SystemsHai “Helen” Li, Duke University
FEATURED SPEAKERS & ABSTRACTS
Keshab K. Parhi
University of Minnesota Twin Cities
Keshab K. Parhi
University of Minnesota Twin Cities
DOMAIN-SPECIFIC ENERGY-EFFICIENT ARTIFICIAL INTELLIGENCE AGENTS
The unprecedented power of large language models holds promise for design of artificial intelligence agents that can demonstrate broad capabilities of intelligence, including reasoning, planning, and the ability to learn from experience... However, the path forward is unsustainable with respect to energy efficiency, growth in data required to train models, and safety and security. We argue that future AI agents will need to be trained using much less data by exploiting prospective learning, as opposed to retrospective. Further reduction in energy is achieved by reducing memory access, sparsity, quantization, mixture of experts, fine tuning, and low-rank approximation.
BIOGRAPHY
Keshab K. Parhi received his Ph.D. from UC Berkeley in 1988. He is the Erwin A. Kelen Chair in Electrical Engineering at the University of Minnesota. He has published over 750 papers, holds 36 patents, and authored the textbook VLSI Digital Signal Processing Systems. Dr. Parhi is the recipient of numerous awards including the 2003 IEEE Kiyo Tomiyasu Technical Field Award and the 2017 Mac Van Valkenburg award. He is a Fellow of IEEE, ACM, AIMBE, AAAS, and the National Academy of Inventors (NAI).
Niraj K. Jha
Princeton University
Niraj K. Jha
Princeton University
BOTTOM-UP DOMAIN-SPECIFIC SUPERINTELLIGENCE
The AI industry is currently focused on achieving general superintelligence in a top-down fashion... This approach is insatiably thirsty for data during training, leading to unsustainable electricity/water costs and CO2 emissions. We propose to take the opposite tack – build domain-specific superintelligence bottom-up, modeled after how superintelligence is achieved in human society. This framework will take inspiration from neuroscience and include episodic and working memories to facilitate reasoning.
BIOGRAPHY
Niraj K. Jha is a Professor of Electrical and Computer Engineering at Princeton University. He is a Fellow of IEEE, ACM, and AAIS. He has co-authored five books and more than 500 papers, including 16 award-winning papers. His research interests include algorithms, architectures, and applications of machine learning and natural language processing.
Naveen Verma
Princeton University and EnCharge AI
Naveen Verma
Princeton University and EnCharge AI
ANALOG IN-MEMORY COMPUTING IS A KEY DRIVER OF ENERGY EFFICIENT AI
After decades of hoping for analog computation, research has brought us to the point of making analog, particularly in-memory computing (IMC), a critical technology for future energy-efficient AI systems. This talk surveys major findings, looking at analog IMC from the lens of practical architectures for delivering scaled-up system-level efficiency inflections.
BIOGRAPHY
Naveen Verma is the Ralph H. and Freda I. Augustine Professor of Electrical and Computer Engineering at Princeton University. His research focuses on advanced sensing and computing systems. He recently co-founded EnCharge AI to commercialize foundational technology developed in his lab. He is the recipient of numerous teaching and research awards.
Tinoosh Mohsenin
Johns Hopkins University
Tinoosh Mohsenin
Johns Hopkins University
PHYSICAL AI AT THE EDGE: CONNECTING PERCEPTION TO ACTION FOR ON-DEVICE AUTONOMOUS INTELLIGENCE
As AI systems move beyond perception to enable real-time, reliable action in the physical world, their computational demands continue to rise. In this talk, I will discuss hardware-aware architectures and the co-design of algorithms and systems that enable efficient, real-time intelligence directly on-device, dynamically balancing computation between edge and cloud.
BIOGRAPHY
Tinoosh Mohsenin is an Associate Professor in the Dept. of Electrical and Computer Engineering at Johns Hopkins University. She directs the Energy-Efficient High-Performance Computing (EEHPC) Lab. Her work has been recognized through over 200 peer-reviewed publications and numerous awards, including the NSF CAREER Award, the Amazon Research Award, and the 2026 Edge AI Foundation Educator of the Year Award.
Krishnan Kailas
Independent Researcher and Entrepreneur
Krishnan Kailas
Independent Researcher and Entrepreneur
BEYOND EFFICIENT ACCELERATORS: SYSTEM DESIGN CHALLENGES FOR SUSTAINABLE AI
The cost of AI computation is becoming impossible to hide. This talk will examine the end-to-end challenges of building energy-efficient AI systems, from software integration and accelerator abstraction to realistic deployment constraints. The talk will argue that approximate computing is an enabling technology for sustainable AI, but only when considered across the full stack.
BIOGRAPHY
Krishnan Kailas is an independent researcher, inventor, and entrepreneur. During his 23 years at IBM T. J. Watson Research Center, he led exploratory systems research projects that helped bring emerging technologies into commercial products. He is a recipient of several IBM Research technical achievement awards and the Mahboob Khan Outstanding Liaison Award from the SRC.
Hai “Helen” Li
Duke University
Hai “Helen” Li
Duke University
HOW MEMORY POWERS INTELLIGENCE: FROM COGNITION TO AI COMPUTING SYSTEMS
As AI continues to scale, modern models increasingly mirror the complexities of human cognition. Yet, a widening gap exists between biological memory and contemporary computing systems. This talk explores this intersection through the lens of memory, reimagining intelligent systems not just as data processors, but as experience-driven learners. By drawing parallels between cognitive science and AI architecture, we highlight novel memory-centric approaches to unlocking adaptive computing systems.
BIOGRAPHY
Hai (Helen) Li is the Marie Foote Reel E’46 Distinguished Professor and Department Chair of Electrical and Computer Engineering at Duke University. Her research includes neuromorphic circuits for brain-inspired computing, machine learning acceleration, and emerging memory design. Dr. Li is a fellow of AAAS, ACM, IEEE, and NAI, and has received numerous accolades including the IEEE Edward J. McCluskey Technical Achievement Award.