In this video from ISC 2019, Jan Mannik from One Stop Systems describes how the company delivers “AI on the Fly.”
The next killer application for High Performance Computing technology with the range and skill set required to deliver effective solutions that not many technology vendors acquire. The demand for these solutions is expanding rapidly with the ever increasing utility of artificial intelligence applications across a wide set of industries.
OSS is a custom manufacturer of specialized high performance computing systems for industries that are showing the largest proliferation of use cases for AI on the Fly systems. These industries include:
- Deep Learning
- Transportation
- Security
- Manufacturing
- Retail
- Media & Entertainment
These edge applications have unique requirements over traditional embedded computing. There is no compromise possible in delivering high performance while maintaining efficient space, weight and power. Delivering the high performance required in edge applications necessitates PCIe interconnectivity providing the fast data highway between high speed processors, NVMe storage and compute accelerators using GPUs or application specific FPGAs. AI on the Fly high performance applications will naturally demand this capability on the edge. Additionally, these solutions often require unique space and power saving form factors and specialized rugged enclosures.
Why is deep learning training only done in the datacenter and inferencing only done in the field when high performance computers that can be deployed in the field exist today? To identify hostile threats from planes or on the battlefield, predicting maintenance requirements at the oil field or piloting autonomous vehicles that learn on-the-fly instead of in a datacenter, these applications require local high performance processing. The need to acquire vast quantities of data at ever faster rates and then apply sophisticated analysis algorithms in real time requires all the traditional capabilities of High Performance Computing but now deployed at the edge and in mobile platforms.
The impact of artificial intelligence is starting to be realized across a broad spectrum of industries. Typically, deep learning (DL) training is a centralized datacenter process and inferencing occurs in the field. To build an AI system, data is collected, run through data scientist training models based on deep learning (DL) frameworks – on the fastest accelerated computers in the world – with the output sent to the field for an “AI at the Edge” system to inference from this model in day-to-day decision making.
AI on the Fly solutions are different than traditional datacenter centric AI infrastructure, as they must deploy in harsh and rugged environments. The solutions must meet unique criteria for shock and vibration, redundancy, large operating temperature, altitude ranges and uninterrupted power sources. Few companies have the expertise across all the technologies required for designing and delivering tightly integrated ‘AI on the Fly’ platforms.”
