Softmare clever programming

Low-code no-code solution

Reservation system as a widget

booking system with no-code current stage - MVP App Development

Algorithmic market stock trading

Machine learning for e-commerce

About artificial neural network

Currrent neural network approach is to much mathematical.

Historical when concept rise and encounts a problem it was solved by adding complexivity. Firs neural networks couldn't go deeper because of linearity. Adding more layers when neuron response function is linear was unfounded. So natural solution was to add nonlinearity but then learnin problem arose. Finaly back propagation aproach moves forward neural networks but we have gradient wanish problem, and so on...

Convolutional Neural Networks adding some indirect layers that detect specific edges to localize objects on images and forward this information to nex layers. In short, it is kind of preprocesing implemented to neural networks itself.

Now many ready solutions, frameworks become popular thanks to simplicity, all complexivity simlified to minimum programing interface.

Neural networks are data storage structue inspired by naural neral system so to bo cosequent more natural aproaches shoud be implemented rather than math. They should work like memory but with generalization skil.

Nature development is based on survviwe and competition.