Applications of Functional Programming To Machine Learning

该项目围绕函数式编程和递归方案领域展开，并探索其与深度学习的关系；本研究旨在研究如何将功能概念与神经网络相结合，并使神经网络的理解受益。我打算重点分析在神经网络的结构和训练中可能存在的功能抽象，以及如何将递归方案作为一种固有特征。这将涉及当前神经网络发展中存在的缺陷，旨在展示如何在统一函数式编程和深度学习领域中实现更好的透明度，代表性和可靠性。该项目将涉及实现一个基于递归方案的神经网络库，同时将其功能形式化为实际的理论理解。

This project revolves around the area of functional programming and recursion schemes and exploring its relationship with deep learning; this intends to research how functional concepts can integrate with and benefit the understanding of neural networks. I intend to focus on analysing the functional abstractions that can exist in the structure and training of neural networks, and how recursion schemes can be found as an inherent characteristic. This will engage with existing flaws in current neural network development, and aims to show how better transparency, representativeness, and reliability can be achieved in unifying the fields of functional programming and deep learning. The project will involve implementing a recursion-scheme based neural network library whilst formalizing its functionality into a practical theoretical understanding.