Upcoming
Amica Conference: Thermodynamic Approaches to Artificial Intelligence. March 9-11, 2026, Karuizawa, Japan.
Amica Conference: Thermodynamic Approaches to Artificial Intelligence. March 9-11, 2026, Karuizawa, Japan.
Notes on the circuit model of quantum computing. Currently mostly an exposition on the menagerie of quantum gates.
We show that natural gradient descent can have a similar computational complexity per iteration to a first-order method, when employing appropriate hardware.
Physics-based Application-Specific Integrated Circuits present a transformative paradigm by directly harnessing intrinsic physical dynamics for computation.
Communications of the ACM article on thermodynamic computing and Normal Computing.
IEEE Spectrum article on thermodynamic computing and Normal Computing.
We present the first continuous-variable thermodynamic computer, which we call the stochastic processing unit (SPU).
We propose a scalable algorithm for employing thermodynamic computers to accelerate a popular second-order optimizer called K-FAC.
Here, we provide a thermodynamic algorithm for exponentiating a real matrix. We describe a simple electrical circuit involving coupled oscillators, which can...
WebLogo 3.9 released with support for python 3.13.
Artificial intelligence wastes an extraordinary amount of energy - but running every computer calculation twice, first forwards and then backwards, could dra...
We consider an alternative physics-based computing paradigm based on classical thermodynamics, to provide a near-term approach to accelerating linear algebra.
We propose electronic analog devices that sample from Bayesian posteriors by realizing Langevin dynamics physically.
Here’s a good review and summary of my book Field Guide to Continuous Probability Distributions : by Eric Regis
Interview on the Future of Computing blog about thermodynamic linear algebra.
Talk at U. Chicago Computations in Science Seminar, 2024-03-20.
We consider an alternative physics-based computing paradigm based on classical thermodynamics, to provide a near-term approach to accelerating linear algebra.
We introduce a method that reduces the overall error from a linear to a quadratic dependence on the imprecision, for Gaussian sampling and linear algebra app...
Wired article on alternative computing approaches, including Normal Computing’s thermodynamic computing.
We consider an alternative physics-based computing paradigm based on classical thermodynamics, to provide a near-term approach to accelerating linear algebra.