School of Computer Science and Statistics
Before fully autonomous driving is achieved, connected autonomous vehicles (CAVs) will operate for a certain period in mixed traffic, which includes both CAVs and human-driven vehicles (HDVs). The dynamic and interactive conditions in mixed traffic scenarios renders CAV decision making particularly challenging. This project will investigate the use of Graph Convolutional Deep Reinforcement Learning for … Read more
Individual fairness concerns the ability of a machine learning model to not being affected in its predictions by one or more sensitive features, such as gender, race, age etc.Recent methods developed techniques for the formal analysis and approximation of fairness in the case of deep Neural Networks (NNs). However such techniques are restricted to simple … Read more
Deep learning has recently achieved state-of-the-art performances in many computer vision classification tasks. In this project we’ll look into developing a tailored deep learning solution based on Neural Network for a problem related to classification with physiological signals. The project will go into details of architecture selection, hyper-parameter learning, validations, and best training practice of … Read more
Biometric authentication refers to systems that verify the identity of a user by unique traits such us fingerprints, voices or particular physiological features (such as ECG which will be the focus of the project). After having developed such a system based on publicly available data, this project will investigate the possibility of bypassing it using … Read more
Deep learning, in particular Neural Networks (NNs), has achieved state-of-the-art results in many applications in the last decade. However, the way these models work and operate is still not fully understood, and in many ways they are approached as black-box when deployed in practice.Unfortunately, this raises several concerns about their suitability to deal with sensitive … Read more
The sudden rise of adversarial examples (i.e., input points intentionally crafted so to trick a model into misprediction) has shown that even state-of-the-art deep learning models can be extremely vulnerable to intelligent attacks. Unfortunately, the fragility of such models makes their deployment in safety-critical real-world applications (e.g., self-driving cars or eHealth) difficult to justify, hence … Read more