Mathematics for Deep Learning
Combining theory, modelling, data, and computation to unlock the next generation of deep learning
Combining theory, modelling, data, and computation to unlock the next generation of deep learning
Machine learning, in particular Deep Learning (DL) based on ‘neural networks’, is one of the fastest growing areas of modern science and technology, and is having an enormous and transformative impact on our lives.
The applications of DL embrace many disciplines including biomedical, physical and social science, computer vision, speech recognition, gaming, music and finance.
The Maths4DL programme focuses on deep learning in areas where reliability, robustness and quantification are key: scientific machine learning.
Alongside the explosive growth in machine learning technology there has been a concern about the lack of understanding behind DL and the way its algorithms make decisions.
This leads to mistrust in the use of some of these algorithms. The huge successes of DL are not all well understood, the results are sometimes mysterious, and there isn’t always a clear link between the data training DL algorithms, and the decisions made by those algorithms.
Maths4DL is putting DL onto a firm mathematical grounding by combining theory, modelling, data and computation to unlock the next generation of deep learning for important applications.