DeepMind: The AI that is revolutionising every industry

The impact of the evolution of technology in the history of human beings is indisputable, in accordance with the experts of the Professional Master’s Degree in Advanced Operating Systems Computing, the impact of technology has facilitated a large number of tasks, directly influencing sectors such as education, work and health, among others.

One of the most complex systems currently on the rise is DeepMind, an artificial intelligence affiliated with Google, which aims to innovate within its own industry.

New era for science by DeepMind

DeepMind combines the best techniques from machine learning and systems neuroscience to build general learning algorithms. Thanks to this tool, it has solved some of the biggest challenges in research; their neural networks have been able to determine the shape of the basic building blocks of life, enabling cures for lethal diseases.

According to research by US biologist Cyrus Levinthal, it was estimated in 1969 that discovering all the possible configurations of a single protein from its amino acid sequence would take at least 14 billion years, more than the time since the origin of the universe. But the AlphaFold project, the technology provided by DeepMind, took just a couple of minutes to solve it. After the training to which it was subjected through video games, scientists challenged DeepMind to solve this enigma, as molecules such as antibodies and hormones compose the tiny cells, essential for their functioning; given the precision of the system, it was able to read the information about its structure, revealing more than 350,000 structures and about 100 million are planned to be published in the following months.

In addition to allowing the deciphering and curing of various diseases, this discovery would ease millions of years of scientific exploration, as well as starting innovative studies to discover new ways of solving other human enigmas.

Creativity based programming

AlphaCode, another instrument coined by DeepMind, can write code at a competitive level; this programme has a level of problem-solving efficiency and creativity comparable to the average human programmer.

The way it works is by copying code from others, a practice similar to that of a programmer and relying on the aforementioned neural networks, supported by deep learning. AlphaCode was trained through other platforms that contain code such as GitHub and Codeforeces; when is exposed to a problem, it will give a large number of solutions, and then choose the best ones.

The system faced programmers and engineers, who were challenged to respond to a series of problems. At the end of the test, its results were in the top 54% of those who did the best, the first time it would achieve such a level. According to researcher Oriol Vinyals, in order to respond to the challenges, it had to combine critical thinking, logic, algorithms, programming and understanding of natural language.

On the other hand, AlphaCode, rather than being competition for a human programmer, Vinyals argues that "many of the tools that have been developed through research, such as compilers, have helped programmers to think at a higher level of abstraction in order to focus their efforts on the function of the program, rather than the code itself", which is why this software tool can become a support both for those specialising in systems and in other disciplines.

Wealth distribution through AI

It is noticeable that DeepMind software can be used to solve various kinds of enigmas, although, it has also been suggested that it would have the ability to contribute to the fulfilment of social objectives to achieve a more just and prosperous society. However, it would not be practical for governments of nations to base their assets exclusively on this.

Researchers at DeepMind developed an approach that combines artificial intelligence with human democratic deliberation to find the best solutions to social dilemmas. Artificial intelligence was able to discover a mechanism to correct the imbalance in the distribution of wealth. The knowledge was acquired through the work of more than 4,000 people and game simulators, which included learning how to distribute public money, aligning values, deciding public policy, and even making legal judgements.

While this development is a great contribution to correcting imbalances in nations suffering from inequality, it has some shortcomings, so it can serve as a support for decision making rather than the basis for everything, prioritising needs, and supplying demands, and moreover, prioritising the most relevant ones, leaving aside those that are not so essential for development.

Although DeepMind, as well as similar devices have facilitated human tasks, and even eliminated the most archaic ones, saving time and resources, it may seem that it will replace the work of man. However, as mentioned above, this system is not perfect, so relying 100% on Artificial Intelligence can be a double-edged sword, therefore, it should be seen as a support for certain activities, or else to have the supervision of an expert.