ANN ARBOR, Mich. Efficient neural networks as a potential solution to energy challenges associated with machine learning.
Training neural networks on traditional computing hardware requires a vast amount of electrical power. This training process is at the heart of all machine learning algorithms and essentially teaches these systems how to operate. Researchers at TRINA are investigating the ability of optical technology to replace these traditional electronic components for advanced computing, particularly in machine learning. In our article, we articulate how optical computing in its many forms can have an impact on industries from short range robotics to parallel computing in data centers.
Link: Weighing in on photonic-based machine learning for automotive mobility | Nature Photonics