The Gatsby Computational Neuroscience Unit at UCL is holding a four-day tutorial on Neural Computation from Tuesday 31st August to Friday 3rd September 1999.

The tutorial is aimed at research students and postdoctoral researchers who are interested in and wish to learn more about neural computation and computational neuroscience. The tutorial is intended for researchers who are trying to understand how the brain computes, but it should also be of value to researchers who are using neural network methods for solving applied problems.

The first part of the tutorial will describe techniques for learning and inference that have their roots in computer science, statistics, physics, engineering and dynamical systems. The second part will describe how these techniques have been applied to understanding computation in the brain.

The tutorial will take place over a four-day period - commencing early afternoon on Tuesday 31st August and concluding mid-afternoon on Friday 3 September 1999.

A summary of topics covered by the tutorial follows.

Day 1 - Afternoon Introduction and tractable models

• Neural networks, autoencoders and generative models
• Mixture of Gaussians clustering and the EM algorithm
• Factor Analysis and Principal components analysis (PCA)
• Mixtures of factor analysers, split and merge algorithms
• Independent components analysis (ICA)

Day 2 - Morning Intractable models and approximations

• Boltzmann machines and sigmoid belief networks
• Gibbs sampling
• Variational approximations and free energy

Day 2 - Afternoon Time series models

• Hidden markov models
• Linear dynamical systems and kalman filters
• Dynamic belief networks
• Sequential importance resampling

Day 3 - Morning Towards neural plausibility

• The wake-sleep algorithm
• Rectified Gaussian belief networks
• Lateral connections and dimensionality reduction
• Learning population codes
• Spiking neurons

Day 3 - Afternoon Introduction to Computational Neuroscience

• Overview
• Statistical and information theoretic treatments of neural coding and decoding

Day 4 - Morning Single cell and network models of computation

• Rate-based network models as non-linear dynamical systems
• Dynamical models of early visual processing

Day 4 - Afternoon Neural plasticity

• Hebb, anti-Hebb and delta learning rules for unsupervised and supervised learning
• Reinforcement learning and neuromodulation

A full programme with session timings will be sent to registered participants prior to the tutorial.

Dr. Geoff Hinton
Dr. Peter Dayan
Dr. Zoubin Ghahramani
Dr. Zhaoping Li
Dr. Hagai Attias
Dr. Sam Roweis
Dr. Emo Todorov
Dr. Carl van Vreeswijk

Venue: The tutorial will be held at the University College London main campus, Gower Street, London, WC1N.

Registration Fee: The tutorial registration fee is £25.00. This fee is non-refundable. Participants will be provided with course papers and light refreshments during the tutorial. Participants are responsible for their own lunch, travel and accommodation costs.

Accommodation: A number of single rooms have been reserved for Gatsby Tutorial participants in one of UCL's halls of residence, Campbell House, 5-10 Taviton street, London, WC1H 0BX. The cost per night is £17.25. Campbell House is a short walk from the UCL campus and is self-catering.  To book a room, please contact: Residence Office, Campbell House, 5-10 Taviton Street, London, WC1H 0BX, Tel: +44 (0)171 391 1479, Fax: +44 (0)171 388 0060. Please quote 'Gatsby Unit' when booking. Guests will be responsible for full payment of their accommodation bill on arrival at Campbell House.

To register for the Tutorial:

Complete and email the Tutorial Registration Form by Friday 20 August 1999

Send a cheque (made payable to University College London)
for the registration fee of £25.00 to: Janice Hankes, Gatsby Unit, UCL, Alexandra House, 17 Queen Square, London, WC1N 3AR

Receipt of your registration form and payment will be acknowledged.

The number of tutorial places is limited.   In the event that the Tutorial is oversubscribed, preference will be given to research students, postdoctoral academic researchers and those who register early.

A final programme, map and other details will be sent out 10 days before the start of the Tutorial.

Contact:  If you need any further information about the Gatsby Tutorial please contact: Janice Hankes janice@gatsby.ucl.ac.uk
Gatsby Unit, University College LondonAlexandra House17 Queen SquareLondon WC1N 3AR,
Tel: +44 (0)171 391 1179 - Fax: +44 (0)171 391 1173.