In this workshop, Dr. Umberto Michelucci, Machine Learning Researcher and Lecturer, will present Autoencoders, an unsupervised learning technique for representation learning. 

Autoencoders are a class of algorithms particularly efficient at extracting interesting features from datasets for a large variety of applications. We will look at what autoencoders are and how to build them with neural networks. We will cover feature extraction, classification/regression with latent feature representation, anomaly detection, and denoising with random and non-random noise. Participants will have the chance to build and train autoencoders for all of these applications with TensorFlow Keras and Python in Jupyter notebooks (bring your own laptop if you wish!). 

Requirements: Basic Python and ML knowledge to understand how neural networks work (layers, loss functions, etc.) is necessary. The hands-on sessions are at different levels to make sure that everyone can profit from the workshop! 

We look forward to seeing you there!

PS: If you are landing here for the first time, to stay up to date about the next workshops join our GDSC community (700+ members) by clicking the 'Join us' button.