Inspiration
We built models and pipelines that could be used to predict the additive manufacturing parameters that is required by the problem, especially in laser-based simulations. The big picture idea was to eliminate the need to spend long hours of running physics-based simulations to achieve results.
This is a Dassault Systemes sponsored project.
What it does
It is a series of functions that takes you through the various stages of predictions of simulations
How I built it
Challenges I ran into
Cleaning the data and gaining insights to build models that are intelligible and interpretable. Reasoning out based on the model performance why the data is structured that way.
Accomplishments that I'm proud of
This is the first time we took up a manufacturing problem and applied ML to help build physical product. We look forward to building more like this!
What I learned
Data Carpentry through and through (~ 14 hours spent just to get the data in a workable way and building models)
What's next for Machine Learning to Build Machines
Build an end-to-end solution for the model
Built With
- automl
- pandas
- python
- scikit-learn
Log in or sign up for Devpost to join the conversation.