Research / Current / Unconventional Computing

I am currently interested in developing software tooling for a range of unconventional computational paradigms that leverage the physical behavior of the underlying hardware substrate to effectively perform computation. Examples of such paradigms include analog computing (with differential equations, CNNs, etc), oscillator-based computing, diffractive optical computing, and quantum computing. A goal of this research is to enable (relatively) easy implementation of programs that can be run on these sorts of computing platforms.

Transmission Line-Inspired Computing

We explore a novel, transmission-line-inspired computational paradigm where an input signal moves through a geometry that transforms, splits, and merges internal signals to perform a computation -- the computational result is then measured from a point in the geometry. Much like a transmission line, the edges in this geometry carry signals from one place to another. However, unlike a transmission line, the edges and nodes in this geometry can implement complex dynamics that significantly change the behavior of the signal. This computing paradigm enables easy prototyping of certain kinds of circuits at a higher level of abstraction: if the edge and node dynamics can be reasonably implemented in the analog domain, then geometries composed of these edges and nodes can also be implemented as an analog circuit.

We are currently using this computational paradigm to rapidly develop analog security primitives that can be effectively implemented in the analog domain. We are in the process of investigating how best to assemble geometries that maximize a set of domain-specific security metrics.

Student: Yu Neng Wang