The third age of machines is near. Quantum and cognitive computers will perform Big Data processing in the cloud. This age follows that of data science, a multidisciplinary science for collecting, processing and enhancing Big Data.
Data science is like petrochemicals for the digital world, refining crude Big into useful, predictive information. Thanks to this, we know in advance the product a customer will choose. To achieve such a feat, data science uses algorithms, and quite extraordinary machine learning can improve these algorithms over time.
1) The current computing model is unsustainable
these algorithms need immense computing power, and giants such as Google, Amazon and Netflix have developed parallel computing systems. The problem is, current equipment is not ideal - a machine is sequential and does not know how to comprehend thousands of pieces of information at once.
We want some kind of super human brain, which has an average of 100 billion neurons with 10,000 connections each. We can wait as microprocessors improve - after all, Moore's Law states that they double their capacity regularly through miniaturization. But we are still far from electronic transistors being as small as a few atoms.
The solution that researchers are considering is to base the treatment of information on a quantum state of small electronic components. According to physicists, in a quantum system a particle can be in two places at one time and an action on one particle can automatically alter another.
In the present state of research, quantum computers have a fault: any unfortunate interaction with their environment changes their super processing power into what is effectively a game of heads or tails. The idea would be to have only a few quantum computers, stored in special ways and designed to interact with all other conventional computers for combined optimization. Such a computer was built by the startup D-Wave, and Google bought it to control its self-driving cars and other technology.
2) Connected objects with a brain
other companies are also getting into the development of super-intelligent machines that are inspired not by quantum physics but by cognitive neuroscience. Their chips have plasticity, that is to say they have the ability to adapt and understand.
These chips, inspired by the human brain, will equip all connected objects eventually to make them more intelligent in their interaction with data and users. This will be the stage of 'isomorphic' machines.
These two revolutions, of quantum and cognitive computing, will begin with very bulky and expensive machines. We believe they will be available as a hosted or cloud service, to allow consumption of their computing capacity on demand, at much lower prices. This is the future.
posted by: TUNJI OLOGUNEBI
No comments:
Post a Comment