SCHEDULE

Agenda is tentative and subject to change. All times are in MDT.

Monday, February 2

5:00 - 7:00pm Welcome Meet Up with light refreshments (Hospitality Suite, La Fonda)

Tuesday, February 3
**Coffee, tea, and water available all day (Santa Fe Room, La Fonda)**

7:00 - 9:00 Breakfast and Badge Pick Up (Santa Fe Room, La Fonda)

8:00 - 8:15 Welcome and Workshop Introduction – Art Edwards, AFRL

8:15 - 8:45 Workshop context setting – Suma Cardwell, SNL

Topic A: Where is AI today and where do we need to go?

9:05 - 9:25 KEYNOTE: Shared experience/ Lifelong Learning Continuous Learning – Ted Senator (Former DARPA Program Manager)

9:25 - 9:35 Neuro-Inspired Continual Learning: What, How, and Why? – Dhireesha Kudithipudi, UT San Antonio

9:35 - 9:45 Efficient inference and learning in recurrent neural circuits – Zeyu Yun, UC Berkeley

9:45 - 10:05 Panel Discussion

Topic B: AI at the edge/ training

10:05 - 10:25 KEYNOTE:

10:25 - 10:35 Understanding and Optimizing Quantization-enabled LLM Serving Systems: A Performance, Energy, and Quality Perspective – Yi Ding, Purdue University

10:35 - 10:45 Spiking Neural Networks Integration in Heterogeneous Architectures with MoSAIC – Farzad Fatollahi-Fard, LBL

10:45 - 10:55 Enable Native Full-Integer Spiking Neural Network Training with Algorithm and Hardware Co-Design – Xiaoxuan Yang, University of Virginia

10:55 - 11:05 Neuromorphic Signal Processing at the Edge: Energy-Efficient SNNs for Audio Processing – Kendric Hood, Kent State University

11:05 - 11:25 Panel Discussion

11:25 - 11:35 *Break before Working Lunch Keynote*

11:35 - 12:35 LUNCH KEYNOTE - Jim Lyke, AFRL/AFOSR

12:35 - 12:50 *Break*

Topic C: Frameworks for fundamental energetic limitations

12:50 - 1:10 KEYNOTE: Energetic Demands of Continual Learning in Biological Brains – Becket Ebitz, University of Montreal

1:10 - 1:20 Development of a Practical Adiabatic Reversible CMOS (ARC) Technology for High-Performance Computing – Mike Frank, VAIRE

1:20 - 1:30 Catherine Schuman, University of Tennessee Knoxville

1:30 - 1:40 Reliable AI in Unreliable Energy-Constrained Distributed Environments – George Michelogiannakis, LBL

1:40 - 1:50 Energy-Efficient Solar Forecasting with a Neuromorphic-Bayesian State Estimation Approach – Kumar Anurag, University of New Mexico

1:50 - 2:10 Panel Discussion

Topic D: Next-Generation Technologies

2:10 - 2:30 KEYNOTE: Quantum Computing Data Centers: Current Trends and Future Outlook – Carleton Coffrin, LANL

2:30 - 2:40 Naïve resource estimation in quantum and brain-inspired computing – Shashank Misra, SNL

2:40 - 2:50 Emergent Spatiotemporal Neuromorphic Computing in Protonic Nickelate Network, Duygu Kuzum, UC San Diego

2:50 - 3:00 Casper: Cascading Hypernetworks- for Scalable Continual Learning – Tej Pandit, UT San Antonio

3:00 - 3:10 Neuro-Symbolic Architectures for Scalable and Energy-Efficient Computation: Opportunities and Open Challenges – Shay Snyder/Maryam Parsa, George Manson University

3:10 - 3:30 Panel Discussion

Topic E: Next-Generation Technologies -Session 2

3:30 - 3:50 KEYNOTE: Edge of Chaos and Entropy Production Minimization as Keys to Neuromorphic Computing – Stan Williams, Texas A&M University

3:50 - 4:00 Energy and Information in Open Thermodynamic Systems – Narayan Srinivasa, Intel

4:00 - 4:10 Advanced packaging for low power, weight and form factor bio-inspired sensing platform – Anastasiia Butko, LBL

4:10 - 4:20 Reshaping LLM Edge Architecture by Emerging Memory Technologies – Ramtin Zand, University of South Carolina

4:20 - 4:30 Neuromorphic Energy-Efficient Multi-Layer Learning with Compartmental Spiking Neurons – Ashish Gautam, ORNL

4:30 - 4:50 Panel Discussion

4:50 - 5:00 Day 1 closeout (Observations and Day 2 Outlook)

5:00 - 5:30 Poster Summary Blitz

5:30 - 7:30 Poster Session and RUMP Session (informal technical discussions) with refreshments (Portal, La Fonda)

A-Graph: A Unified System Representation for Automated Cross-Stack Optimization - Di Wu

Catwalk Neuron for Efficient Temporal Neural Networks - Di Wu

Control of analog brain dynamics & learning as a computational principle for neuromorphic computing derived from the brains SWaP optimized multi-scale co-design principles - Kris Bouchard

Denoising with Equilibrium Propagation - Siddharth Mansingh

Efficient prediction in cultured neurons and humans - Sarah Marzen

Energy-Efficient Acceleration of Transformers with Stochastic Photonic Computing - Ishan Thakkar

Federated neuromorphic computing in distributed edge environments - Seung-Hwan Lim

Genesis: A Spiking Neuromorphic Accelerator with On-chip Continual Learning - Vedant Karia

GPU Support and Other Improvements for SuperNeuroMAT - Prasanna Date

Hybrid Splat-and-Mesh Simulation for Rapid Humanoid Training in Isaac Sim - Wei Xu

In-DRAM Stochastic Computing for Energy-Efficient Acceleration of Transformer Neural Networks - Ishan Thakkar

Inverse-Designed Photonic Dot-Product Primitives for Energy-Efficient Neuromorphic Computing – Raktim Sarma

Reliable Cryogenic Circuits in Unreliable Environments - George Michelogiannakis

Sample- and Modality‐Efficient Spatiotemporal Learning for Neuroscience and Robotics - Xihaier Luo

Sense-as-You-Go: A Neuromorphic Framework for Efficient Edge Sensing and Processing - Di Wu, Mohsen Rakhshan

Signal propagation dynamics across the Drosophila hemi-brain connectome modeled as a multiplex networks reveal a parallel-hierarchical (‘U-Net’) sensory-cogntive-motor architecture - Kris Bouchard

Spatial Mental Modeling from Limited Views - Manling Li

UQ-VarQA: Benchmarking and Characterizing NISQ Computers Through Uncertainty Quantification of Variational Quantum Algorithms - Qiang Guan

Wednesday, February 4

**Coffee, tea, and water available all day (Santa Fe Room, La Fonda)**

7:00 - 9:00 Breakfast and Badge Pick Up (Santa Fe Room, La Fonda)

8:30 - 8:45 Day 2 Welcome

Topic F: Neuromorphic Primitives and Circuit Design

8:45 - 9:05 KEYNOTE: Jennifer Hasler, Georgia Tech

9:05 - 9:15 Time-Domain Neuromorphic Computing Circuits – Cory Merkel, Rochester Institute of Technology

9:15 - 9:25 Biological Plausibility and Energy Efficiency in Artificially Intelligent Learning Hardware – Joseph Friedman, UT Dallas

9:25 - 9:35 Efficient Log-Domain Circuit Realization of Biologically-Plausible Neurons – Abdulkarim Alorf, Duke University

9:35 - 9:45 Biologically-Inspired Multiscale Neuromorphic Architecture – Christian O’Reilly, University of South Carolina

9:45 - 10:05 Panel Discussion

Topic G: Novel Use-Cases for Emerging Devices

10:05 - 10:25 KEYNOTE: Integration of Efficient Sensing, Encoding, and Learning via Memristive Dynamics – Joshua Yang, University of Southern California

10:25 - 10:35 Memristor-Based Spiking Neural Network Accelerator for Neuromorphic Computing – Chenyun Pan, UT Arlington

10:35 - 10:45 Energy-efficient and Scalable Inferencing with Hardware Innovation – Beyond-CMOS Technologies – Guojing Cong, ORNL

10:45 - 10:55 Energy-Efficient Temporal Processing with Artificial Spin Ice Based Neuromorphic Nonlinear Vector Autoregression Machine – Ishan Thakkar, University of Kentucky

10:55 - 11:05 Hybrid Opto-electronic Integrated Encoders for Ultralow Power Data Compression – Nicholas Nobile, SNL

11:05 - 11:25 Panel Discussion

11:25 - 11:35 *Break before Working Lunch Keynote*

11:35-12:35 LUNCH KEYNOTE: Spiking Manifesto – Eugene Izhikevich, CEO of SpikeCore, Chairman of Brain Corp

12:35 - 12:50 *Break*

Topic H: Beyond CMOS/ Emerging Devices and Technologies

12:50 - 1:10 KEYNOTE: Emerging devices for Neuromorphic Computing – Yuping Zeng, University of Delaware

1:10 - 1:20 Modeling Ferroelectric and Antiferroelectric Capacitors with Physics-Informed LSTMs – Nicholas Ramos/Hai Li, Duke University

1:20 - 1:30 Van der Waals ferroelectric tunnel junctions for next-generation logic-in-memory hardware – Cheng Gong, University of Maryland

1:30 - 1:40 First-Principles Investigation of the Resistive Switching Energetics in Monolayer MoS2: Insights into Metal Diffusion and Adsorption – Jameela Fatheema, UT Austin

1:40 - 1:50 SuperNeuroABM: Scalable, Biologically Realistic, and Hardware-Aware Neuromorphic Simulator – Xi Zhang, ORNL

1:50 - 2:10 Panel Discussion

Topic I: Energy Consequence of Biological Systems

2:10 - 2:30 KEYNOTE: From Dendrites to Waves: Multiplexed Codes for Energy-Efficient Cortical Information Processing – Krishna Jayant, PhD, Purdue University

2:30 - 2:40 A Theoretical Framework for Time, Space, and Energy Scaling in Neuromorphic Systems – Brad Theilman/ Brad Aimone, SNL

2:40 - 2:50 Large Multimodal Brain Models for Neuroscience and Computing – David Park, BNL

2:50 - 3:10

3:10 - 3:30 Panel Discussion

Topic J: Advances in Neuroscience relevant to Emerging Computing

Session Chair: Dr. Chou Hung, Army Research Laboratories

3:30 - 3:50 KEYNOTE: Overview DOE Neuromorphic Workshop – Garr Kenyon, LANL

3:50 - 4:00 Energy efficiency of neuromorphic architectures for Generative AI – Frances Chance, SNL

4:00 - 4:10 Autonomous Learning of Attractors for Neuromorphic Computing – Riley Acker, LANL

4:10 - 4:20 Compositional associative memories support efficient, robust, and flexible computations – Christopher Kymn, UC Berkeley

4:20 - 4:30 Mapping Challenges of Connectomes for Performant Neuromorphic Simulation – Felix Wang, SNL

4:30 - 4:50 Panel Discussion

4:50 - 5:00 Day 2 closeout

5:00 - 7:00 ECI Workshop Dinner (Room TBD, La Fonda)

Thursday, February 5

**Coffee, tea, and water available all day (Santa Fe Room, La Fonda)**

7:00 - 9:00 Breakfast and Badge Pick Up (Santa Fe Room, La Fonda)

8:30 - 8:45 Day 3 Welcome

Topic K: Energy Aspects of Quantum Computing

Session Chair: Erik DeBenedictus, Zettaflops LLC

8:45 - 9:05 KEYNOTE: Quantum Computing Benchmarking Initiative – Mike Descour, SNL

9:05 - 9:15 The footprint of Quantum Computing: Technologies and Trade-offs – Kasia Kryzyzanowska, LANL

9:15 - 9:25 EnerQy: Energy Estimation for Quantum Computing – Di Wu, University of Central Florida

9:25 - 9:35 Energy-Aware Orchestration for Hybrid Quantum-Classical Computing: Towards Unified Telemetry and Cost Models – Ozgur Ozan Kilic, BNL

9:35 - 9:45 Quantum Enhanced Josephson Junction Field-Effect Transistors for Low-Energy Power-Efficient Microelectronics Applications – Wei Pan, SNL

9:45 - 10:05 Panel Discussion

Topic L: Thermodynamic Computing

Session Chair: Chiping Li, AFOSR

10:05 - 10:10 Introduction to Thermodynamic Computing - Chiping Li, AFOSR

10:10 - 10:30 KEYNOTE: From Turbulence-in-a-Box to an Exploding Star: Energy Cost of Modern Computational Fluid Dynamics – Alexei Poludnenko, University of Connecticut

10:30 - 10:40 Scalable Ising Solvers for Combinatorial Optimization – Shimeng Yu, Georgia Institute of Technology

10:40 - 10:50 A Digital Twin of Future Quantum HPC Systems – Priyrabrata Senapati, Kent State University

10:50 - 11:00 Kinetic Inductors Enable Reversible Logic – Erik DeBenedictis, Zettaflops LLC

11:00 - 11:10

11:10 - 11:30 Panel Discussion

11:30 - 11:45 *Break before Working Lunch*

11:45 - 12:35 FUNDING AGENCIES PANEL, LUNCH

Robinson Pino (DOE ASCR)
Hal Greenwald (AFOSR)
Chiping Li (AFOSR)
Jim Lyke (Latin American Office for AFOSR)

12:35 - 12:50 *Break*

Topic M: Macro- and micro solutions to HPC Energy budgets

12:50 - 1:10 KEYNOTE: Mapping neuro-inspired principles and motifs to on-chip AI in sensor networks and signal processing applications – Angel Yanguas-Gil, ANL

1:10 - 1:20 Chiplets already fuel the Exascale. Chiplet modularity is key for the next 100-1000x energy savings – Patricia Gonzalez-Guerrero, BNL

1:20 - 1:30 Thermal Management Opportunities from Exploding Energy Demands in Datacenters – Foluso Ladeinde, Stony Brook University

1:30 - 1:40

1:40 - 1:50 Energy Consequences of Machine Learning Inference on HPCs – Mohammed Atif, BNL

1:50 - 2:10 Panel Discussion

Topic N: Co-Design

2:10 - 2:30 KEYNOTE: A Full-Stack Neuromorphic System Implementation for Development and Benchmarking – Rajit Manohar, Yale University

2:30 - 2:40 Towards Efficient and Scalable Graph Solutions with Neuromorphic Learning – Shruti Kulkarni, ORNL

2:40 - 2:50 Fault-Tolerance as a Basis for Energy-Efficient Co-Design – Hemanth Kolla, SNL

2:50 - 3:00 A principled procedure for designing brain-derived SWaP optimized neuronal units for low-power neuromorphic analog computation and digital communication – Kris Bouchard, LBL

3:00 - 3:10 Predictive Quantum Simulations for State-of-the-Art and Beyond CMOS Device Technologies: Enabling the Full Stack Co-design of Novel Computing Systems – Juan Mendez, SNL

3:10 - 3:30 Panel Discussion