The Limited Times

Now you can see non-English news...

IA and 5G, a double-strand bond

2020-08-04T12:19:18.792Z


Next generation networks will enable pervasive use of artificial intelligence. Meanwhile, researchers from Huawei's Italy Research Center in Segrate resort to machine learning to improve performance (ANSA)


by Alessio Jacona *

"In this laboratory we are working today on solutions, tools and materials that will be ready in ten years, in time for the 6G debut". Renato Lombardi, director of Huawei's Italy Research Center in Segrate (Mi), one of the most important laboratories in the world that the Chinese giant dedicates to research on high frequency wireless technologies for 5g applications, is a man constantly projected into the future.

Today the world is awaiting the debut of the fifth generation networks, while Lombardi finally sees the fruit of studies, tests and research started two years ago. Sitting in the conference room, the director speaks quickly as someone who has to maximize the time available, while he tells us that 5g is one of the basic technologies for the future, because it enables and strengthens many others by connecting them together with connections characterized by very low latency, high speed and ultra-wide band.

All this is also true for artificial intelligence, which in the case of 5G reveals a double-strand bond: on the one hand, the pervasiveness of a super efficient network will allow it to be used more and more widely and articulated, for example to operate a increasing number of connected devices, which interact to perform increasingly complex tasks, from managing the production line in a factory to driving a self-driving car.

On the other hand, AI plays and will play an increasingly important role in optimizing the functioning of new generation networks. «With machine learning we can solve problems that before we faced only by using complex mathematical models» confirms Lombardi.

The issue is that the transmission systems, to effectively cover the territory with the signal, must overcome a long series of obstacles: from the simple interferences that are created between antennas, to their positioning on the territory (conditioned by many variables), from the obstacles present on the range of action (buildings, trees, statues), to the substantial unpredictability of the movements of each user.

If we add to this that the 5G antennas work with "beamforming", that is, they do not limit themselves to irradiating an area as was the case for 4G, but instead orient and concentrate the signal so that it reaches and follows the individual user (to maximize the quality and minimize the intensity), then it is clear how difficult it is to set the transmission systems and their thousands of parameters a priori.

Paradoxically, the solutions being developed at the Italy Research Center (as well as some already in use) provide for the use of less intelligence and more "brute force", because they replace complex mathematical models that try to predict every possible situation a priori with machine learning and neural networks, which instead manage transmission systems by adapting parameters in real time to changing needs.

"Machine learning and deep learning are technologies that become transversal influencing every area - explains Lombardi - when it comes to radio systems, their application does not concern so much the improvement of performance, but the simplification of the operation, because they allow for example to use less complex and expensive hardware systems, as well as consuming less energy ».

Talents and skills are needed to help build the future of telecommunications between ten years and more. To find in quantity, the Huawei research center - whose Italian and non-Chinese management is a more unique than rare case for the company - has developed a large number of collaborations with our universities, in which it invests to create Joint Labs and 5-year research programs. Siena, Pavia, Polytechnic of Milan and Polytechnic of Turin, University of Ferrara and University of Naples are just some of the realities involved in the collaborations.

Some of the young researchers involved in the program have the opportunity to work at Huawei, for example in the laboratory directed by Lombardi: «The funniest thing is to see that the new hires, who belong to the generation of so-called millennials, have an approach to solving of the problems that are already deeply imbued with machine learning today », says Renato Lombardi. "For them it is now an indispensable technology - he concludes - so much that they seem more interested in understanding how a system is trained than in understanding how it solves the problems that are submitted to it".


* Journalist, innovation expert and curator of the ANSA.it Artificial Intelligence Observatory

Source: ansa

All news articles on 2020-08-04

You may like

News/Politics 2024-03-25T15:04:44.045Z
News/Politics 2024-03-13T08:14:00.611Z

Trends 24h

Latest

© Communities 2019 - Privacy

The information on this site is from external sources that are not under our control.
The inclusion of any links does not necessarily imply a recommendation or endorse the views expressed within them.