At the NVIDIA GTC, one of the main subjects is Artificial Intelligence and how it will improve our life. As usual, NVIDIA is always striving for performance and perfection, making the AIs possible with their fast GPUs
First of all, NVIDIA is a graphics processing chip manufacturer getting most of its revenue out of selling GPUs (Graphics Processing Unit). These are world-class performance, performing in competitive gaming, professional visualization, and cryptocurrency mining.
The NVIDIA GTC, which stands for GPU Technology Conference, is an annual congress where developers, buyers, and innovators attend. Let’s find out more about their progress on AI.
NVIDIA Getting Involved In-Vehicle Performance
A car has been for ages, just a steering wheel and four wheels. This aspect changed when humanity started looking for environmentally-friendly transportation.
The cars and the trucks of the future will no longer need just an optimized AI for autonomous driving, but also for monitoring the driver. In the future, we will use a car more like we use Siri, having a conversation with it. The car will become the co-pilot, not just the vehicle.
Last year, NVIDIA announced their involvement in this industry, by presenting the NVIDIA DRIVE AGX Orion. While most of the current electric cars use a system that does one trillion operations per second, the NVIDIA’s AGX Orion does nearly 7 times more.
These numbers could help humans to reach things that seemed impossible before. The AGX Orion will be implemented in the production of cars starting in 2022.
No Traffic, Less Time Waiting in Intersections
We stick to the vehicle subject by presenting the NoTraffic, a computer vision researcher, presented last year at GTC, that runs on the NVIDIA Jetson platform. The NoTraffic uses AI sensors at intersections to analyze traffic and optimize traffic lights. NoTraffic comes to contrast with the traditional traffic detectors that use just and inductive-loop.
This computer can maximize or minimize a light’s duration to maximize the cars’ performance, minimize the duration of waiting, and minimize the pollution of CO2. NoTraffic classifies all road users, from buses to bicycles
It can also calculate the collision percentage by calculating the speed, acceleration, and directions, so it can infer when two cars will probably collide. It reduces delays of an estimated 2,700 hours and prevents 33 tons of carbon emissions per intersection.
Building AI For Social Good
For traditional human rights investigations, researchers need to move to an area, have interviews, go to the crime scene. Then, they have to check the court records and collect records from the hospital.
They concluded that sometimes, this cannot be possible. That is why the program Human Rights Watch started including AI strictly for investigations.
The AI will provide its investigation by satellite and drone data. Furthermore, it offers access to analyses of public datasets, and it will use social media posts and photos. This could be very viable when the researchers can not reach a conflict zone or a closed country.
Human Rights Watch also uses two NVIDIA DGX Stations, which were provided by NVIDIA back in 2018. Its deep learning possibility could provide in the future alerts directly to the Human Rights and humanity agencies when they detect a potential human rights abuse.
Tracking Cancer Made Easy
At NVIDIA GTC, Faisal Mahmood presented an AI Software, made to help the doctors to determine whether a patient has cancer.
A study shows that one out of four breast cancer cases remains indeterminate when a biopsy is used to diagnose it. The diagnosing of biopsies is very time-consuming; one biopsy could take at least 15 minutes.
Here comes the AI that uses four on-prem machines with NVIDIA GPU clusters. The machine-learning will go to two possible outcomes: one where it will analyze all the biopsy, and a doctor will just check the AI work or another one where the software will separate the negative biopsies from the rest, and the doctor will just check the biopsies left.
This will have a huge impact on the health and well-being of patients.
All summed up, at the NVIDIA GTC, we were presented one of the most life-enriching AIs. From improving the autonomous cars, spending much less time in traffic, fighting for human rights, to even tracking cancer, these show us what AI could bring to us in the future.