Artificial Intelligence: What To Expect For Your Business In 2020
By: Evandro Barros.
“The destruction of the past – or rather the social mechanisms that links our contemporary experience to the earlier generations, is one of the most characteristic and eerie phenomena of the late twentieth century. Most young men and women at the century’s end grow up in a continuous present without any organic relation to the public past of the times in which they live.”
This is a passage from the book “Age of extremes” from Eric Hobsbawm, famous historian that wrote a series of books about how the eras are influenced by the facts that define them, despite the calendar, questioning the end of the XIX century and the beginning of the XX century.
On the Artificial Intelligence’s world, the emergence from techniques as deep learning, and the evolution of reinforcement marked the beginning of a new era ongoing. Do the second decade from the XXI century has begun in 2014 or 2007 with the iPhone launch? To find the marks, it is necessary to take a distance from the facts.
The 2020s will be decisive for AI’s startups and companies that are working with this technology internally. The evolution of some architectures capable of learning with a low volume of data can expand to markets still impounded that suffer because of limited techniques that require a high amount of data.
On the other hand, quantum computing can promote significant advancement in parameterization techniques, complex situations and algorithm evolution. More than learning with low volume, the algorithms will be capable of learning from zero information.
The techniques presented by the Deepmind in 2017 (A general reinforcement learning algorithm that masters chess, shogi and Go through self-play) are just a demonstration of what is coming. Quantum computing will be using deep learning, not only in board games but using DNA and molecular combinations to simulate nature. The advantage of that is generating data through reverse engineering, promoting the learning of the machines and the discovery of new combinations.
The natural language processing or the capability of the machine’s understanding and communication with humans will be another decisive element in this decade, a watershed.
According to Statista.com, there are more than 100 million users of virtual assistants in the U.S. This number increased by more than 38% in 2019; the more significant was Alexa from Amazon. The growth in the use of the assistants can signal a drastic transformation in the market. As the iPhone has started the App industry, enabling the emergence of billionaire companies in the 2010s, the virtual assistants will change our relationship with the Internet, search systems, books, education, traffic, and especially the way that society perceives an intelligent machine.
There are more than 30 projects in AI in several companies, and we can say that the biggest challenge in this field is making theories, studies and scientific experiences, in reality, which is something extremely complex.
This will be the biggest challenge in the 2020s, very similar to what happened in the years followed by the internet’s creation. Companies will connect their departments through Artificial Intelligence, not isolated experiences and without context, as happened in the last decade. It will also be when we see the results from the first big bets in technology; that companies that initiate the process but did not show the power of their achievements yet. When this happens, nothing else will be like before.
The future has already begun, be prepared!
Evandro Barros is a historian, music producer, professional former trader working with derivatives, currently mastering in Math at USP, in São Paulo, Brazil. He studies the application of advanced deep learning. He is a serial entrepreneur, founder of I2A2 – Institut d’intelligence artificielle appliqué, co-founder of DATA H – Artificial Intelligence Inc.
Background cover image via Rawpixel (Royalty Free Photo)
Disclaimer: The views and opinions expressed in this article are those of the authors and do not necessarily reflect the official view of the FCBB.