4:15 PM - 4:45 PM
[19p-232-6] Physical Reservoir Computing by optical delay feedback systems
Keywords:optical delay feedback system, reservoir computing
Reservoir computing is an architecture of recurrent neural networks composed of input, reservoir and readout layers. The remarkable feature of reservoir computing is that the weights in the reservoir and input layer are fixed by random numbers and only weights of readout layer is trained. In addition, reservoir computing can be implemented by optical delay feedback systems. In this talk, we will clarify the operational principle of optical delay feedback systems based on the simplified Lang-Kobayashi equation.