Several companies in the world and few scientific teams are working on using live neurons for the construction of biological, intelligent computers. Such processors would be incomparably more efficient than silicon chips serving artificial intelligence – said PAP expert in the field of neuroscience Dr. Ewelina Kurtys.
The living brain is extremely efficient and efficient in terms of information processing – if it was used in technology, this would mean a revolution in computer science – believes the expert.
“The main reason why we work on computers from live neurons is that Nerve cells are very energy efficient. We know this even from observing the human brain. At this stage, we see this the biggest advantage of using live neurons. Current research on artificial neural networks focus mainly on speed and precision, and less attention is paid to energy savings. Meanwhile, its consumption by artificial intelligence increases exponentially. In the future it will be a significant problem, “explained Dr. Ewelina Kurtys, a researcher working at the Swiss company Finalspark.
As she emphasized, we are just observing the beginning of work on live computers.
“This is a very innovative field. In addition to FinalSpark, only two startups are working on it – in the USA and Australia. When it comes to scientific research, only a few work is carried out. This is particularly about research on the use of human neurons, because only for about 15 years we can relatively easily receive them from stem cells made of human cell cells,” the researcher noted.
She added that creating a “live processor” is a huge challenge for many reasons. “First of all, neurons are cells that are very sensitive to any changes in the environment. It is necessary to closely monitor all parameters, such as access to oxygen, nutrients, temperature, etc. In our experiments, we create a group of about 10,000 neurons that create a structure about 0.5 mm. We grow them on a plate with a suitable medium,” she reported.
Although nerve cells are very sensitive, they can work for a long time in appropriate conditions.
“We know that neurons in the human brain can live up to a hundred years – most of them operate from birth to death. Theoretically, therefore, lively, bred neuron networks can also work for years. We keep our cells alive for three months, because we need so much time at the current stage of research,” said Dr. Kurtys.
Live networks are affected by traditional electronics. “The group of neurons we grow is currently connected to eight electrodes. With their help, we send and receive electrical signals from the cells. Neurons communicate using chemical and electrical impulses. In this way we train our network and recover information from it” – described the specialist.
She added that in the laboratory where she works, you can even do it remotely. “We have developed such a system during a pandemic. Thanks to this, we opened our laboratory for other teams from different parts of the world. From the proposals of experiments sent to us, we choose the most interesting and allow other scientists to implement them on our networks for free. We also have the first commercial clients who rent our laboratory remotely for experiments,” said Dr. Kurtys.
As she decided, the programming of live cells is a powerful challenge: “The process of learning live neural networks is still the subject of intensive research. In fact, no one knows exactly how the brain learns; which are behind this mechanisms; what are brain learning algorithms. And this is a huge scientific challenge. People are trying to conduct research on this at different levels. For now, we conduct experiments in which we send different signals in electric and we measure the answers.
It turns out that a living network can work a bit like the so -called Black box. “Our goal, however, is not really understanding how neurons work, how they learn. What is really interested in us is the result. So if we manage to achieve learning and the ability to perform various sensible tasks, even if we don't know exactly how neurons do it, it's also fine” – explained the specialist.
She added that it should also be possible to transfer data between the networks: “When it comes to transferring information from one network to another, we believe that this will also be found. If we develop techniques for reading information from the network and training them, we will probably be able to transfer information between the networks in this way – read the content of one network and train the other on this basis, i.e. transfer information to it.”
Slightly different studies are also conducted in the world – on the use of systems already developed by nature, e.g. nervous systems of simple organisms, which are even insects.
“However, we are going in a different direction. We want to create a bioprocessor based on human neurons. In the case of small cell groups, a type of neuron may not be so significant – we could, for example, use rat or mouse cells. However, with time, when these structures are getting bigger, it can gain meaning. We know that the human brain is exceptional and has the greatest calculation options among the brains of all species. The use of human cells will in the future to process more complex algorithms.
In the field of learning the researcher and her colleagues, they only achieved their first successes. “At the current stage, and it is very early, we were able to save one bit of information on the living network. This is really not much, but the success can be considered that we were able to control the network consisting of live cells. Our competition from the Australian company Cortical Labs managed to train a similar network to control the game 'pong' – said Dr. Kurtys.
However, this is the beginning of the road – she reminded. “I think that obtaining more complex algorithms is a matter of time and financial outlays. For example, we are looking for investors, because our project is completely financed by the company's founders.
Meanwhile, there are already questions about the ethical aspects of such works; Even whether complicated neuron structures will be aware.
“This is a very wide topic, and all theses are difficult to prove. The very issue of consciousness is very abstract and is based solely on opinions, not on facts. However, these questions are important. We are also involved Is ethics, which is why we want specialists in these areas.
If lively computers are created and widely used, they are unlikely to cause traditional processors to disappear. “I see a future in which various technologies will be used simultaneously, for different purposes. We will be dealing with a variety of processors. Numerous systems for specific applications are already being created. In addition, in addition to ordinary computers, quantum processors are developed, which are to be perfect for selected tasks. Similarly, it will be with live neural networks. A traditional computer, for example, is better than a man with repetitive tasks, but a lot better with a man Filters information that all technologies will find their place on the other hand.
Dr. Ewelina Kurtys is a scientist working on biological computers at the Swiss company Finalspark. She is an expert in the field of neuroscience and the commercialization of new technologies and strategies for their development.
Marek Matacz (PAP)
mat/ bar/ mhr/
