Could this chip out-compose the likes of Vivaldi and Bob Dylan?

The answer to that must surely be no. Vivaldi wrote more than 500 concertos for mandolin, cello, flute, viola, recorder and lute, and around 230 of them were for violin. I doubt whether even a self-learning neuromorpic microchip could be that prolific despite it already demonstrating some ability to compose music.

The chip is based on resistive switching metal oxide RAM (OxRAM) technology which is seen as an emerging memory technology that is attracting considerable interest because of its high speed, high density and low cost of fabrication characteristics.

Shown at ITF2017 the device is said to be the world’s first self-learning neuromorphic chip that is inspired by the way the human brain works.

We all know the human brain is unquestionably the most complex and sophisticated entity known to computer and electronics designers so what neurological elements caught the imagination of the creators of the neuromorpic chip?

Apparently, the key attributes were the human brains phenomenal computing power and the miniscule amount of power it requires to achieve monumental results.

Researchers at Imec, the nanoelectronics and digital technologies organisation, are attempting to combine cutting-edge hardware and software to create chips that feature the human characteristics of self-learning. The organisation’s long-term ambition is to develop the process technology and related building blocks to create artificial intelligence systems that are sufficiently energy efficient they can be integrated into sensors. It is their belief that such intelligent sensors will drive the Internet of Things forward. This would not only allow machine learning to be present in all sensors but also allow on-field learning capability to further improve the learning.

According to Imec by optimising hardware and the software, the neuromorphic chip features learning and intelligence characteristics on a small area, while consuming only very little power. The chip is self-learning, meaning it can makes associations between what it has experienced and what it experiences. The more it experiences, the stronger the connections will be.

As an example of what may prove possible in the future is if neuromorphic chips are integrated into sensors for health monitoring they could identify a particular heart rate change that could cause heart abnormalities. They could also learn to recognise slightly different ECG patterns that vary between individuals. Such neuromorphic chips might enable more customised and patient-centric monitoring.

Now this I see as having some huge advantages but as for music composition I would need a lot of convincing that music written by a memory device would really Please Please Me or could it actually be a case of The Times They Are A Changing?

Click the learn more links below to find out more.

Can this help embedded engineers keep pace with IoT challenges?

Binary hyper-drive. It’s computing but not as we know it captain