is there any way to train a model to learn a matrix and all variables in it then generate same matrix

actually, my supervisor told me to find a way to computer learn the matrix which I created for my thesis and the computer can make the same matrix with the same attitude as my matrix

anyone know how can I do this? deep learning? neural network? is there anyway, can somebody give me that way!?

I made my matrix in python with sympy

Answer 1

[Sam Chak]

Hi @hessam asadi

The answer is generally yes, it is possible to train so long as it is a non-singular (invertible) matrix.

First, generate the data based on your desired Attitude Representation and Transformation Matrices.

Next, follow the examples in this article: Deep Learning in MATLAB.

https://www.mathworks.com/help/deeplearning/ug/deep-learning-in-matlab.html

Then, either export the Deep Learning network to TensorFlow, or export the network to ONNX model format.

For more info, please check out this article: Interoperability Between Deep Learning Toolbox, TensorFlow, PyTorch, and ONNX.

https://www.mathworks.com/help/deeplearning/ug/interoperability-between-deep-learning-toolbox-tensorflow-pytorch-and-onnx.html

[hessam asadi]

hey @Sam Chak

thank you for you answer , i need to know which one of the Work flow in this article can solve my problem https://www.mathworks.com/help/deeplearning/ug/deep-learning-in-matlab.html