Meet Startup @TW

Kneron shows its leadership in Edge AI Summit 2019


In late November, Kneron, an edge AI solutions provider, showcased its progress and vision at the 2019 Edge AI Summit. In his speech, Kneron’s CEO and Taiwanese native, Dr. Albert Liu, PhD, detailed his vision of the Edge AI Net made possible through their unique reconfigurable technology.

The Edge AI Summit is in its second year and brings together venture capitalists, industry press and analysts, and edge AI companies to foster dialogue and spur growth between vested parties in this growing segment within the AI industry. Even with speeches from the likes of Google, BMW, Microsoft, Facebook and NVIDIA, Dr. Liu’s stood out for its focus on the vision that Kneron has for the future of Edge AI.

More Wall-E than Skynet, Kneron’s vision is to build an Edge AI Net that allows intelligent devices to be able to infer and communicate with each other without going through a centralized AI brain, usually found in the cloud. In effect, this would democratize AI inferencing and distribute its power away from a centralized brain.

The key to realizing this vision is Kneron’s reconfigurable technology. Due to the way it’s built, Kneron’s Neural Processing Unit (NPU) technology is able to not only switch between audio, voice, and 2D/3D vision recognition in near real-time, but can also adapt to all major AI frameworks and CNN models in the market.

This allows Kneron’s hardware and software solutions to enable any device with the ability to communicate and make actionable decisions based on changing conditions regardless of AI framework and without going to a central cloud-based brain. Latency is significantly reduced while security and privacy are improved.

Dr. Albert Liu stated, “It was exciting to open eyes and ears at the Summit when we shared our vision of the Edge AI Net, and that there are already products in the market from major brands that are powered by Kneron edge AI solutions.”

The future looks bright for Kneron as they pioneer the way forward for the Edge AI Net. If they succeed, our future will have more Wall-E’s and less Skynets.