Neural networks learning differential data

In many of machine learning problems, it is essential to use not only the t raining data, but also a priori knowledge about how the world is constraine d. In many cases, such knowledge is given in the forms of constraints on di fferential data or more specifically partial differential equations (FDEs). Neural networks with capabilities to learn differential data can take adva ntage of such knowledge and easily incorporate such constraints into the le arning of training value data. In this paper, we report a structure, an alg orithm, and results of experiments on neural networks learning differential data.