COMSM0075

Week 1 - CH - information theory

• lecture 1: introduction to the unit, motivation, definition of entropy and some of its properties. This corresponds to, roughly, the first four pages of "notes1".
• notes1 lecture notes for information theory pdf
• lecture 2: more on information theory, the source coding theorem, joint and conditional entropy, mutual information and the information processing inequality.
• old online lecture motivating Shannon's entropy youtube / slides (pdf)
• old online lecture: Shannon's entropy youtube / slides (pdf)
• old online lecture: joint and conditional entropy youtube / slides (pdf)
• old online lecture: mutual information youtube / slides (pdf)
• old online lecture: the data processing inequality youtube / slides (pdf)
• worksheet 1 (based on week 1). pdf

Week 2 - CH - information theory in the brain

• lecture 3: recap of mutual information, information and the brain.
• lecture 4: this was a problem class but you were also asked to watch the differential entropy lecture and read the notes.
• lecture notes for information in the brain pdf
• lecture notes for differential entropy pdf
• lecture 6 information in the brain youtube / slides (pdf)
• lecture 7 differential entropy examples youtube / slides (pdf)
• lecture 8 differential entropy and Shannon's entropy youtube / slides (pdf)
• lecture 9 mutual information for continuous probabilities youtube / slides (pdf)

Week 3 - CH - the Bayesian Brain / Kalman filters

• lecture 5: the Bayesian brain.
• lecture 6: more on the Bayesian brain, start of Kalman filtering.
• for a gentle revision of Bayes's rule see: coms10011 note on Bayes
• lecture notes revising Bayes's rule pdf
• lecture notes on Bayesian fusion in the brain pdf
• lecture notes on the Kalman filter pdf
• note that infomax wasn't covered this year

Week 4 - RPC - Neural circuits and learning

• L1: Neural circuits and learning: introduction Stream / slides (pdf)
• L2: Supervised learning and backprop Stream / slides (pdf)
• Guest lecture (Charles Kind): Introduction to neuromorphic computing | slides (pdf) / video

Week 5 - RPC - Neural circuits and learning

• Lab 1: Supervised learning and backprop (link; solutions)
• L3: Visual system: deep learning? Stream / slides (pdf)
• L4: Reinforcement learning Stream / slides (pdf)
• Guest lecture (Sean Froudist-Walsh): Introduction to cognitive computational neuroscience | slides (pdf)

Week 6 - Reading week

Week 7 - RPC - Neural circuits and learning

• Lab 1+2: Reinforcement learning (Lab 2: link; solutions)
• L5: Unsupervised learning Stream / slides (pdf)
• L6: Temporal processing and RNNs Stream / slides (pdf)

Week 8 - RPC - Neural circuits and learning

• Lab 2: Reinforcement learning
• L7: Gated RNNs, Brain vs Maquina Stream / slides (pdf)

Week 9 - 11 Coursework weeks