Why Deep Solve
Inspired by the need for solving complex differential problems, AI and deep learning can lead the way to powerful and efficient computational research. The problem with deep learning is the steep learning curve to write code, tune network parameters and learn AI fundamentals. Deep Solve provides an easy-to-use and code-free interface for solving differential equations with the power of AI.
What it does
Deep Solve leverages AI to solve complicated equations in a fraction of the time that traditional solvers take. Also, in the real world, not all parameters of a differential equation are known. Unlike traditional solvers, Deep Solve can analyze provided data to predict the value of these parameters while simultaneously solving the differential equation.
Deep Solve uses a Physics Informed Neural Neural Network to learn the solution to a system of differential equations. By embedding the differential equations into the loss function of a neural network, the network can be trained to approximate the equation's solution.
How we built it
Deep Solve was build using DeepXDE, a library for Physics-Informed Neural Networks built on TensorFlow. We used Anvil to create the interface, and Google Colab to perform the computation.
What's next for Deep Solve
The next steps for Deep Solve include creating an even more robust interface, with example equations and instructions to help people get started solving equations faster.
Log in or sign up for Devpost to join the conversation.