Exploiting Defects in 2D Materials for Stochastic Computing Applications
Abstract
In this proposal, we aim to exploit and engineer defects in two-dimensional (2D) materials to develop energy and area efficient hardware primitives tailored for stochastic computing applications. Recent breakthroughs in the manipulation of 2D materials, such as graphene and transition metal dichalcogenides (TMDs), have revealed that their intrinsic defects can significantly alter electronic properties, presenting a unique opportunity for hardware innovation. Our research will focus on three primary objectives: (1) employ advanced microscopy, spectroscopy, and electrical characterization techniques to map and understand the electronic changes caused bydefects in 2D materials; (2) development of a scalable fabrication process to engineer these defects for the realization of stochastic hardware primitives; and (3) integration of these stochastic computing (SC) primitives for acceleration of neural networks, optimization algorithms, and cryptographic applications. For example, stochasticspiking neural networks (SSNNs) capture the inherent randomness of synaptic transmission, thus improving the modeling of neural behavior and processes involved in learning and inference. For Ising spin problems, commonly found in physics and optimization, stochastic computing facilitates the simulation of random thermal fluctuations, essential for studying phase transitions and solving combinatorial problems. Additionally, the inherent randomness of stochastic primitives is crucial for true random number generators (TRNGs), which are vital for enhancing security in computing and hardware applications. Overall, this project is expected to lead to the development of a novel class of hardware for stochastic and probabilistic computing. These advancements are anticipated to deliver enhanced performance, lower power consumption, and a more compact design, meeting the increasing needs for sustainable and secure computing architectures for future DoD applications.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Nov 08, 2024
- Source ID
- N000142412565
Entities
People
- Saptarshi Das
Organizations
- Office of Naval Research
- Pennsylvania State University
- United States Navy