Tree-like Memory in Hardware
Modulating proton distribution under high speed electric pulses in strongly correlated perovskite nickelates
Running artificial intelligence can require a mind-boggling amount of energy. Of course, this is not likely in all cases, yet advanced applications can really consume a lot of energy – as such the imperative to make the way these algorithms run more energy-efficient is quite considerable.
Therefore, any kind of progress made in materials can have a significant effect on the field of artificial intelligence. That in turn may mean that it could have interesting consequences for computing.
A recent article in Nature talks of Perovskite neural trees.
They experiment with material as a proof-of-concept compatibility with existing semiconductor platforms.
“We report experimental realization of tree-like conductance states at room temperature in strongly correlated perovskite nickelates by modulating proton distribution under high speed electric pulses. This demonstration represents physical realization of ultrametric trees, a concept from number theory applied to the study of spin glasses in physics that inspired early neural network theory dating almost forty years ago. We apply the tree-like memory features in spiking neural networks to demonstrate high fidelity object recognition, and in future can open new directions for neuromorphic computing and artificial intelligence.”
As such it is about the material: “strongly correlated perovskite nickelates.”
Perovskite is a calcium titanium oxide mineral composed of calcium titanate.
I am not sure to what extent the progress it yet, but apparently they have been able to demonstrate this material can conduct differently.
“We demonstrate that electric pulses, as fast as tens of nano seconds, are effective in perturbing the proton distribution in the nickelate lattice and can tune its resistivity in a systematic fashion enabling ultrametric tree-like conductance states”
Their team could be one of the first to demonstrate: “…artificial “tree-like” memory in a piece of potential hardware at room temperature.” (SciTech Daily, 7th of May 2020)
It seems researchers in the past have only been able to observe this kind of memory in hardware at temperatures that are too low for electronic devices.
According to Hai-Tian Zhang, a Lillian Gilbreth postdoctoral fellow in Purdue’s College of Engineering (SciTech Daily, 7th of May 2020):
“Mimicking these features in hardware is potentially interesting for brain-inspired computing.”
Potentially does however mean that it is still in the works.
Yet it provides an interesting window into what future hardware within the field of artificial intelligence could be.
This is #500daysofAI and you are reading article 349. I am writing one new article about or related to artificial intelligence every day for 500 days. My focus for day 300–400 is about AI, hardware and the climate crisis.
