NM500, An New Approach to Image Recognition at SZOIL
Written by: Jesus Vasquez | Edit by: Jin Gao
On Saturday July 6, 2018, Shenzhen Open Innovation Lab, SZOIL, hosted an open workshop on ‘Image Classification’ conducted by University of Texas at Austin Undergraduate students Jesus Vasquez and Jin Gao majoring in Computer Science and Electrical Engineering, respectively.
The workshop was structured in a way to slowly and effectively ease participants into this new topic by first establishing fundamental knowledge on Artificial Intelligence. The first hour of the workshop was dedicated to learning basic yet important concepts about AI and Machine learning, as well as presenting some practical and real world applications of this technology.
This includes an exposure to Artificial Neural Networks as well as a demonstration of how these computational networks are often utilized in most AI Image applications today. Finally, we introduced participants to ‘NM500’, a new hardware board highly optimized for Machine Learning.
Contrary to conventional AI image recognition algorithms, the learning neural network on the NM500 is implemented by using over 570 hardware neurons. After the groundwork had been laid with participants, we continued on to the second hour of the workshop, in which we utilized a friendly tool named ‘Knowledge Studio’ that allows anyone, regardless of technical expertise, to create their own model for learning. Participants were asked to use this tool to create a simple Facial Recognition model, with guidance from the instructors available at every step. In concluding the workshop, all participants were successful in doing so.
Overall, the workshop was a success with the supports from Vicky Xie, Ben Chen, and David Li. We are thankful for all the participants’ cooperation,Tsing-Hua Graduate Campus Research Lab, and their great interest in exploring the new NM500.
More info on NM500: http://theneuromorphic.com/nm500
Github: https://github.com/nepes-ai
