IBM is developing a pre-processing unit

   259

It is expected that such development will largely benefit the prosperous industry of AI. According to the mathematician of IBM Zurich Thomas Parnell, they have developed a general solution for the AI ​​learning process with a 10 times speedup.

According to the research laboratory, the main purpose of the invention is to increase the speed with which AI machines can learn new information.

A block that uses mathematical duality to filter important pieces of information in a data stream and ignores non-essential one, is mainly intended for large data machine learning.

Despite the great potential, there is still a lot of work to improve the unit and before it can be ready for commercial use. The company plans to continue developing the system in the IBM’s cloud, where it will be called the Duality-Gap-based Heterogeneous Learning.

The company can benefit greatly from this discovery, given the high current trend in the artificial intelligence industry. However, it remains to be seen whether the company will be able to improve this new technology and how it will operate in the real world in the near future.