According to experts in robotics and artificial intelligence (AI), we are close to tremendous breakthroughs in the field. . A featured article published in the Journal of Economic Perspectives argues that this revolution expected to take place soon in the field of robotics that could be compared with what the Cambrian Explosion has been in terms of biological life on Earth.
Life on earth experienced about half a billion years ago a short period of very rapid diversification called by specialists the "Cambrian Explosion." Many theories have tried to explain the cause of the Cambrian Explosion. Among them, one of the most provocative considers that the evolution of vision allowed animals to dramatically increase their ability to find mates and hunt.
The present technological developments on several fronts are creating the grounds for a similar explosion in the adaptability and diversification of robotics. Many of the actual hardware technologies on which robots depend on particular communications, data storage and computing, have been improving at rates of exponential growth.
For instance, two emergent technologies, "Deep Learning" and "Cloud Robotics", could lead to another cycle of explosive growth. In Cloud Robotics, every robot learns from the experiences of all robots. As the number of robots grows, this could lead to rapid growth of robot competence.
Deep Learning algorithms are used as a method of learning for robots. By using Deep Leaning, robots can generalize their associations based on very large "training sets" that typically include millions of examples that are often cloud-based.
According to experts, one of the robotic capabilities that has been recently enabled by these combined technologies is vision, which interestingly is the exact the same capability that may have played a major role in the biologic Cambrian Explosion.
These circumstances lead to idea that a Cambrian Explosion of robotics may soon occur. How soon it is still hard to tell. Some experts argue that we should consider the history of computer chess, where heuristic algorithms search and brute computing force can beat now the best human player, yet no specialized chess-playing software knows inherently how to handle even a simple adjacent problem.
According to this view, specialized robots might improve at performing well-defined tasks. However, in the real world, the problems yet to be solved overpass the presently known ways to solve them.
But Deep Learning algorithms of today, unlike computer chess programs, use general learning techniques with general domain structure. They can be applied to a range of perception problems, like vision and speech recognition. Taking this into consideration, it is reasonable to assume that in not too distant future robots will be able to perform any associative memory problem at same levels as a human.
With the use of Deep Learning algorithms, even high-dimensional inputs problems could be solved by machines. Unlike computer chess, where improvements have gradually occurred at expected rates, the improvement of Deep Leaning is very fast.
It seems likely that Deep Learning will become able soon to replicate the performance of many of the perceptual parts of the brain. Similar methods may also replicate cognitive functions since the architectures of the cognitive and perceptual parts of the brain appear to be anatomically similar. These advances in AI and robotics give experts a reason to believe that Deep Learning techniques may someday lead to artificial cognition.
The timing of tipping points is yet difficult to predict, but according to experts it seems certain that in the near future will occur an explosion in robotics capabilities. Commercial investment in robotics and autonomy has significantly accelerated, especially in autonomous cars.
Automotive companies, as well as high-profile tech firms like Google, Apple, Amazon and Uber, are announcing significant projects in this area. Key technologies are contributing to the present excitement in the robotics field and an explosive revolution is frenetically anticipated by experts in the field.
