As the amazing progress enabled by the continued miniaturization of the field-effect-transistor slows down, there is developing interest in non-traditional computing as a path to energy-efficiency and increased functionality. In this talk, I will describe one such path based on the concept of probabilistic or p-bits that can be scalably built with present-day technology used in magnetic memory devices.

I will argue that p-bits can be robustly placed in between deterministic bits that are at the heart of digital computers and quantum bits that are at the heart of quantum computers. I will show that these p-bits can be used as building blocks for constructing autonomous p-circuits that can accelerate many current applications like optimization, invertible logic and machine learning algorithms, while providing a bridge to the Noisy-Intermediate-Scale Quantum (NISQ) era quantum computers. Along this direction, I will describe a recent experimental demonstration of an 8-bit p-computer implementing a quantum-inspired optimization algorithm.

An underlying theme throughout this work is an Atoms to Systems approach that that includes emerging materials and novel phenomena (e.g. spintronics, quantum materials), transport theory and device physics, circuit simulation, system-level behavioral synthesis (e.g. using FPGAs), and an understanding of Machine Learning and Quantum Computing algorithms that drive the search for new and specialized devices.

Kerem received his PhD in Electrical and Computer Engineering from Purdue University in 2015, where he continued on as a post-doctoral researcher between 2015 and 2020, before joining the department of Electrical and Computer Engineering in UC Santa Barbara in 2020.

His PhD work established a modular approach to connect a growing set of emerging materials and phenomena to circuits and systems, a framework that has also been adopted by others. In his postdoctoral work, he used this approach to establish the concept of p-bits and p-circuits as a bridge between classical and quantum circuits to design efficient, domain-specific hardware accelerators in the new, beyond-More era of electronics.

Kerem's work has been published in many refereed journals and conferences including Nature, Nature Electronics, Science Advances, Physical Review X. He has delivered more than a dozen invited talks in international conferences and workshops, including American Physical Society (APS) March Meeting in 2016, IEEE Device Research Conference (DRC) in 2017, Magnetism and Magnetic Materials (MMM) Conference in 2017, the IEEE International Electron Devices Meeting (IEDM) in 2019 and the International Conference on VLSI Design (VLSID) in 2020.

Kerem served on the technical program committee for Design, Automation and Test in Europe Conference (DATE) in 2020 and in 2021. He has also served on the technical program committee of the IEEE Conference on Rebooting Computing (ICRC) 2020.

Researchers should cite this work as follows:

• Kerem Yunus Camsari (2020), "Probabilistic Computing: From Materials and Devices to Circuits and Systems," https://nanohub.org/resources/34113.

  BibTex | EndNote

1. adiabatic quantum computing
2. devices
3. machine learning
4. p-bits
5. probabilistic computing
6. quantum computing
7. quantum materials
8. spintronics