Applied Deep Learning

Deep learning is reshaping what is possible with data — from image recognition and natural language processing to complex prediction tasks that were previously out of reach. This course equips participants with the foundational concepts and hands-on skills needed to design, train, and improve neural network models using TensorFlow and Keras. From understanding the theoretical principles of neural networks to leveraging pre-trained architectures, participants gain the practical knowledge to apply deep learning confidently in real-world settings.

While Applied Deep Learning can be taken as a standalone course, it is also part of EPFL’s learning pathway in applied machine learning. It is the fourth of four Microcredentials leading to the Certificate of Open Studies (COS) in Applied Machine Learning https://www.formation-continue-unil-epfl.ch/formation/applied-machine-learning/, providing a flexible pathway toward an EPFL credential in applied machine learning.

Identify supervised learning scenarios as regression or classification tasks. Construct suitable feature spaces through feature engineering, selection, and data preprocessing. Explain the theoretical foundations of core machine learning concepts including gradient descent and the bias-variance trade-off Implement end-to-end machine learning pipelines for classification and regression tasks using Scikit-learn Evaluate a range of supervised learning algorithms, including linear models, k-nearest neighbors, decision trees, support vector machines, and ensemble methods, using appropriate metrics. Diagnose model failures such as overfitting, underfitting, and ill-conditioning through hyperparameter tuning and regularization. Build neural network architectures using Keras and TensorFlow, selecting the appropriate model for a given data type. Leverage pre-trained models to improve generalization and address tasks with limited datasets.

Nongraded

Data and technical professionals who already use Python for data analysis and machine learning and want to develop deeper expertise in neural networks and deep learning. It includes data and quantitative analysts seeking to introduce more sophisticated machine learning models into their everyday workflows, as well as research scientists aiming to design and apply advanced neural network models. It also appeals to software engineers moving into machine learning and AI who want to build practical, real-world applications.