The mathematical/computational study of the brain shares a rich history with machine learning, with substantial mutual interactions between the two fields. In this course, students of machine learning will learn some of that history and how contemporary thinking in neuroscience can inform and be informed by developments in machine learning. The course will cover some basic biological background, an overview of brain models, and go in depth into models of learning in brains and machines. It will also outline trends in neuromorphic computing, the design of brain-inspired computational hardware.

At the end of this module you will be able to:
1. Compare and contrast the underlying aims, objectives, and methodologies employed in neuroscience and machine learning.
2. Argue how developments in neuroscience and machine learning have mutually influenced each other.
3. Decide on the application of techniques in computational neural modelling.
4. Derive mathematical relationships between theories of learning in brains and machines.
5. Compose and modify training algorithms for biologically realistic "spiking neural networks".
6. Assess potential future points of cross-over between machine learning and neuroscience.

- History of links between neuroscience and machine learning
- Challenges for machine learning (robustness, sample efficiency, etc.)
- Basic biology of the brain (neurons, synapses, spikes, etc.)
- Basic models of the brain
- Different factors and constraints on learning in brain and machines
- Biological and machine learning rules and their mathematical relationships
- Training spiking neural networks with surrogate gradient descent and other algorithms
- Overview of Different types of neuromorphic computing hardware
- Open issues in neuroscience and their relationship to machine learning