Links to and summaries of recent publications:

Modeling Performance of Data Collection Systems for High-Energy Physics: Computing through multiple, specialized components is one approch to continue offering increased performance per watt as Moore’s Law continues to slow down. But when you put one of these new components in a large system - such as the data acqusition system at the LHC which processes terabytes of data per second - how does this affect how much power you’ll save, how much data you’ll process, and the quality of that data? We develop a framework to quantitatively estimate these trade-offs and estimate the impacts of multiple potential changes to data processing at the LHC.

Hyperdimensional Computing Provides a Programming Paradigm for Oscillatory Systems: How can devices which oscillate - anything from a pendulum to a photon - be used to compute? We demonstrate how oscillators can be linked in circuits to compute multiple applications including graph compression, factorization, and neural networks driven by continuous, time-varying analog signals.

Deep Phasor Networks: Demonstrated deep neural networks which utilize a phasor-based activation function. These networks can compete with traditional networks on common image-recognition tasks, and carry the advantage that they can inherently be executed via a spiking mode without any conversion step between modes. Code

Bridge Networks: Applied concepts from vector-symbolic architectures to create a network which learns the relationship between two domains of information. E.g. given an image, it can predict a label, and when given a label, it can predict an image. The unique architecture used also allows the network to perform self-distillation and exhibit potential continual learning capabilities. Code

A Dual-Memory Architecture for Reinforcement Learning on Neuromorphic Platforms: Designed and implemented a model for reinforcement learning implemented on the Intel Loihi neuromorphic chip. This model utilized two ‘memory’ segments - one slow, one fast - to enable spike-based reinforcement learning entirely on-chip. The tasks of the multi-arm bandit, a maze, and blackjack were demonstrated. Code

Stochasticity and Robustness in Spiking Neural Networks: Investigated the use of stochasticity in training spiking neural networks and its impact on the robustness of these networks to weight perturbation. Demonstrated that networks trained under stochastic conditions are likelier to have improved robustness.

Cellular Memristive-Output Reservoir (CMOR): Implemented a cellular-automata based reservoir system in a 14-nm CMOS process. Tested resulting integrated circuits and demonstrated that a memristor-based readout of the reservoir was able to carry out a non-linear classification problem on the inputs (XOR).