Forget supercomputers or the latest technological innovation: one of the most impressive machines in the world is right there in your head.

While computers might be great at carrying out relatively simple steps at super-fast speeds, the human brain is still much more sophisticated when it comes to advanced tasks like pattern recognition and creative thinking.

That’s why scientists have for a long time been attempting to mutate the human brain. And they’ve just made a breakthrough.

Fielding questions in an article published by the World Economic Forum, Angeliki Pantazi, member of a team of scientists who worked on the project in conjunction with colleagues from IBM Research, Zurich, said today’s computers were based on the von Neumann architecture, in which the processing chip and the memory are physically separated.

This architecture is highly inefficient for the big data requirements of cognitive computing, in which large amounts of data must be transferred between memory and the computational units at high speeds. To build efficient cognitive computers, which can learn on the fly, we need to transition to non-von Neumann architectures in which memory and logic coexist in some form.

Neuromorphic computing is a very promising approach to non-von Neumann computing that draws inspiration from the inner workings of the biological brain.

What is it about the human brain that inspires scientists to want to mutate it?

The human brain is the world’s most energy-efficient computer. It combines processing and memory together and can solve complex challenges using just 20 watts of energy. By mimicking the way biological neurons function, both energy and areal costs can be drastically decreased for complex computational tasks, such as pattern recognition, feature extraction and mining of data in noisy environments.

Artificial neurons and synapses (the part of the nervous system that allows neurons to pass signals to one another) are computationally very powerful: already a single artificial neuron can be used to detect patterns and discover correlations in real-time streams of event-based data. For example, in the internet of things, sensors can collect and analyse large volumes of weather data collected at the edge for faster forecasts. Artificial neurons could also detect patterns in financial transactions to find discrepancies or use data from social media to discover emerging cultural trends in real time.

This development has been described as groundbreaking. How so?

The results are a fundamental contribution to the field of neuromorphic engineering that revolutionizes the way data is represented and processed. As such, they are a scientific contribution in an active field of research, and it will be at least a couple of years before they find a way to an end-user product. One application where this technology could make a difference is in the area of “data on the edge”, which requires powerful sensors in remote locations to both collect and process data with minimal energy requirements.

Where next with this research?

Only a few months after our paper was accepted in Nature Nanotechnology, we published the next step in our work. In this paper we developed a spiking neural network computing architecture using phase-change memristors to mimic both the synapses and neurons in the neural system for the first time. We also demonstrated enhanced pattern and correlation distinguishing capabilities using a level-tuning mechanism inspired by features found in the human auditory cortex.

Source: World Economic Forum

Author: Yemi Olarinre