This coding exercises aim at the acquisition of practical experience with the most used tools in the field by:
- Application of deep learning techniques to small dataset
- Application of model selection and validation techniques on simulated and real datasets
Image classification task over a dataset of plants using Convolutional Neural Networks models. The dataset we have worked with is composed of
3542 images of plants divided in 8 different classes. Because of the small size of the dataset, our main experiments
consist of transfer learning and data augmentation techniques. Into the following section we’ll describe our work
and the decisions that led us to the CNN model that performed the best accuracy result (0.86%) over the test dataset.
Full Report
Multivariate time series classification project in which several artificial neural network models
were used. The obtained results were compared and analysed to identify the most performant model for this type of
problem.
Full Report