UCLAIS Challenge 2: Neural Networks
Learn more about neural networks through this UCLAIS ML tutorial series challenge on predicting the alcohol content of wine from physicochemical test results.
Loading...
Learn more about neural networks through this UCLAIS ML tutorial series challenge on predicting the alcohol content of wine from physicochemical test results.
In this challenge, you will be exploring how to train neural networks and then use them for inference. In the previous UCLAIS challenge on premier league match result prediction, you worked on a classification task, whereas in this challenge, you will be working on a regression task, where instead of trying to predict a class from some set of discrete classes, your model will output some continuous value.
This is an excellent opportunity to experiment with a range of neural networks techniques using PyTorch. To get started, check out the Jupyter notebook at the link below:
https://github.com/UCLAIS/ml-tutorials-season-4/blob/main/doxa-challenges/challenge-2/getting-started.ipynb
For this challenge, you will use the Wine Quality Dataset, but instead of predicting the quality of the wine, you will be predicting the alcohol content percentage of the wine based on a set of other physicochemical attributes. More information is given in the Jupyter notebook above.
While this is an excellent opportunity to explore how to use neural networks, other approaches (such as the classical machine learning models covered earlier in the tutorial series) may work as well, so feel free to experiment! If you take any interesting approaches, we would like to hear about it on the DOXA Community Discord server.