Rostelecom published the fourth annual monitoring of global digitalization trends. For the second year in a row, artificial intelligence has been the most notable trend.
Experts note a steady growth in the visibility of the technology across all data sources - approximately 30% per year. With reference to the MIT study, the authors claim that the development of AI is approaching its limit due to the limited computing power, but this problem can be solved by using quantum computers.
Several promising technologies in AI highlighted by experts:
Federated learning - a machine learning technique on multiple decentralized edge devices or servers with local data samples;
Neural architecture search - automation of neural networks training with minimal human participation;
Machine reading comprehension - machine reading algorithms capable of reading and understanding text and then answering questions about it.
Also, the autonomous vehicles trend has been growing rapidly over the past 5 years. Experts note a significant growth in the market for technologies for autonomous vehicles (VW Group and Ford collaboration for creating unmanned taxis and delivery, the expansion of partnerships between Uber and Toyota, active investment by large corporations in this area), etc. According to the authors of the study, the main barriers to industrial use are infrastructure and regulation.
Ashmanov Neural Networks