When disaster strikes, the clock ticks faster. Civilians need to be evacuated. Supplies need to be moved. First responders need to be deployed. Delays can cost lives.
AI will soon help those on the ground keep up amidst the chaos, NVIDIA Vice President for Accelerated Computing Ian Buck explained Tuesday, detailing a partnership between NVIDIA and Carnegie Mellon University.
“This is an area where I think we can make a huge impact saving lives,” Buck told an audience of more than 2,000 policymakers, technologists and entrepreneurs gathered in Washington for GTC DC.
The goal of the humanitarian aid/disaster relief initiative is to create AI for the fast-paced scenarios where the right information at the right time can save lives, Buck explained.
The effort will bring together the best minds at CMU and NVIDIA with a “build it, break it, then make it better” design cycle to create tools for disaster response.
CMU brings the experience of over 200 faculty working on a broad spectrum of topics within the field of AI. NVIDIA brings thousands of engineers, an international reputation for R&D and the industry’s leading AI platform.
Solving ‘Impossible’ Problems
These teams will bring world-class technical skills not only to solve so-called “impossible” problems, but to develop new methods for problem-solving, Buck explained.
Nothing’s off the table. The effort will begin with basic science, incorporate new and existing algorithms, and make use commercial off-the-shelf components to jointly deploy AI solutions.
The broad-based partnership will include work in five key areas:
- Autonomy – The effort will create autonomous, unmanned aerial, aquatic and terrestrial systems able to operate in the unstructured spaces encountered during disaster scenarios.
- Human-AI Interaction – The partnership will create new kinds of interfaces, such as virtual, language-based assistants for victims sheltering in place, and assistants able to aid decision makers trying to interpret large amounts of data in real time.
- Sensors, Hardware, Cloud and Edge Compute – Meshing sensors and computing power on devices in disaster zones with the cloud promises to help first responders move faster.
- Hardware – CMU will aim to develop an industry-leading GPU-based capability to address the most challenging AI- and accelerated computing-related problems for those working on humanitarian assistance and disaster recovery.
- Developer Training – NVIDIA’s Deep Learning Institute will help train CMU’s team to apply AI to humanitarian aid/disaster relief.