How does NVIDIA, at the forefront of AI technology, predict the future of AI?

Semiconductor maker NVIDIA is actively developing not only hardware but also software such as machine learning and artificial intelligence (AI). The vice presidents of each division of NVIDIA are predicting 'how AI will be applied in the future'.

2022 Predictions on How AI Will Impact Global Industries | NVIDIA Blog
https://blogs.nvidia.com/blog/2021/12/07/2022-predictions-ai-global-industries/

◆ Brian Katanzaro (Vice President, Applied Deep Learning Research)
Katanzaro predicted in 2020 that 'conversational AI will be used to intervene in real-time gameplay and flesh out a character-driven approach to enhance the immersiveness of the game.' That. However, in reality, it was only applied to business, not games, Katanzaro said.

'Companies will compete for the introduction of new conversational AI tools to enable them to work more efficiently and effectively using natural language processing. Speech synthesis will be human in 2022. It's supposed to be as emotional and compelling as voice, and I think it will help industries such as retail, banking and healthcare to better understand customers and provide better service. ' predict. Companies will also use AI to understand the context and emotions of what customers say, Katanzaro said. It also says it will accelerate the adoption of tools like 'Github Copilot' to help programmers work more efficiently.

'GitHub Copilot', a function that automatically complements the 'continuation' of the source code, has appeared on GitHub, with the cooperation of OpenAI --GIGAZINE

◆ Sara Tariq (Vice President of Automotive)
'The days of programmable cars and the loss of value for once-purchased cars are just around the corner. Build a software-defined architecture that can afford to support new applications and services with automatic wireless updates. So, more and more car makers will be trying to renew their driving experience. Over time, cars will become better and safer. '

In addition, shifting to a software-defined approach will help eliminate stress and hassle in daily driving, Tariq said. In addition to the AI assistant becoming a personal concierge, vehicle occupants will always have access to services provided by AI, and real-time conversational AI will be able to use real-time conversational AI to alert and control the vehicle. I'm expecting it. In addition, automakers will be actively investing in simulation and training of deep neural network models to validate more end-to-end stacks, allowing them to be designed from a long-term perspective. Mr. Tariq said.

◆ Rev Lever Redian (Vice President of Simulation Technology)
'The 3D standard for expressing virtual worlds will evolve,' Leva Redian asserts. 'Developers, businesses and individual users will contribute to the foundation of open standards, similar to the dawn of the Internet. Standards such as Universal Scene Description and glTF are Web 3.0 and digital twins . It will evolve rapidly to meet the underlying demand. '

'In 2022, there will be an explosive increase in the use of data generated from virtual worlds using accurate physics simulators to train advanced neural networks,' Leva Redian said. increase.

◆ Kimberly Powell (Vice President and General Manager of Healthcare)
Powell predicts that AI will dramatically improve drug discovery technology. For example, DeepMind, an AI company under the umbrella of Google, has made an open source of the algorithm for protein three-dimensional structure analysis 'AlphaFold', which has dramatically increased the structural analysis data of proteins, and the purification of chemical substances by AI will bring new drugs. It states that the chances of discovery have increased one million times.

Artificial intelligence company DeepMind releases protein structure analysis algorithm 'AlphaFold' as open source, making it available to anyone --GIGAZINE

by OIST

In addition, AI is expected to increase the number of SaaS (Software as a Service) type medical devices and software, reduce costs, and improve automation and accessibility. By not only providing the hardware, but also upgrading remotely, medical institutions will be able to continue using the device for a long time after it is installed.

◆ Ronnie Bashista (Senior Vice President of Communications)
Bashista said that 5G (5th generation mobile communication system) combined with AI technology will provide network slicing that allocates dedicated communication bandwidth to specific applications, ultra-low latency in wireless communication environments, security and isolation. It is stated that the improvement of the above can be realized. In addition to personal communications, Bashista predicted that new use cases would be created for factory automation, robot management, automotive systems, logistics systems, and more.

◆ Kevin Diering (Senior Vice President of Networking)
“A data center is a new unit of computing. Applications that used to run on a single computer can no longer fit in one box. In the new computing world, software definition and hardware acceleration It's becoming more and more important, 'says Diering. As the use of AI progresses, networks will become like highways connecting many servers that function as computers, and data center-scale cloud computing will be required by connecting software-defined data processing devices. increase.

In addition, Mr. Diering said that the idea of 'zero trust ' that does not unconditionally trust users, terminals, areas, etc. has become commonplace in security, and everyone connected to the corporate system is authenticated and malicious. An attacker will have to constantly monitor for intrusion into the network. “Data is encrypted using IPsec or TLS , and each node needs to be protected by advanced routers and firewalls,” Diering said.