Gradient optimization method using exponential damping and second-order discrete derivative for neural networks and multidimensional real functions
This project focuses on the development of a gradient-based optimization method for multidimensional real functions and neural networks using exponential decay with Tensorflow and Keras.
- Tensorflow
- Keras
- Matplotlib
- NumPy
- Pandas
- Optuna
- Optimization of multidimensional real functions
- Optimization of neural networks
- Utilizes exponential decay for optimization
- Integration with Tensorflow and Keras for seamless implementation