[17p-WL1_301-1~13] Neuromorphic computing at the convergence of biology, electronics, and photonics

4:20 PM - 4:45 PM JST(7:20 AM - 7:45 AM UTC)

[17p-WL1_301-8] Linearization of Neural Networks Using Koopman Operators and Its Applications

〇Jun Ohkubo¹ (1.Saitama Univ.)

Keywords:

Koopman operator,optical circuit,data-driven technique

In this talk, I will first review recent developments of data-driven techniques with the Koopman operator. Then, I will discuss the applications of approximate linear embedding techniques for nonlinear information processing in hardware capable of exploiting linearity, such as optical circuits. Specifically, I will introduce attempts at linear analysis using the Koopman operator for neural networks that incorporate nonlinear information processing.