Lecturer: Professor Kim Plunkett (firstname.lastname@example.org)
Demonstrator: Dr. M. Duta (email@example.com)
Relation to Advanced Options: Specific materials covered in this Block Practical are not covered in any Advanced Option.
Overview: Connectionist modelling work exploits artificial neural networks that mimic some of the basic properties of neural processing in the brain. Densely connected webs of simple processing units propagate and transform complex patterns of activity. When exposed to a training environment, these networks undergo a learning process. The study of the dynamics of these systems and their learning capabilities may provide us with important clues as to the nature of the mechanisms underlying human cognition and its development in infants and young children.
In each class, participants work through a series of self-paced exercises (see the Handbook) under the supervision of the course convenor and course demonstrator. The main objective of the exercises is to introduce the user to some of the essential computation properties of different types of neural networks, and to offer the user an opportunity to replicate and develop some well-known connectionist models of human cognition. A brief introduction to connectionism is available here.
Learning outcomes: Having attended the Block Practical, students should be able to understand the working principle of connectionist networks and should be able to design simple modelling experiments.
Deliverables: Students are expected to write up their modelling project as a formal report not exceeding 3000 words to be handed in to the Examination Schools by the published deadline, and a further copy should be included in the portfolio of practical work to be submitted to the Examination Schools by Friday of 10th Week Hilary Term.
Timetable and estimated workload: Friday afternoons 2.00 pm to 5.00 pm during Weeks 1, 2 and 4 and Wednesday afternoon in Week 3 of Michaelmas Term 2020. 12 hours on practical itself, plus an additional estimated 10 hours on reviewing the literature and writing report. Participants will be encouraged to engage is some preparatory reading before attending the practical class (see Chapter 1 and Chapter 2 of the Handbook) and homework in between the practical classes.
|Timetable for Michaelmas Term 2020|
|1||16th October||Learning to use the simulator||Chapter 3|
|2||23rd October||Learning internal representations||Chapter 4|
|3||28th October||Auto-encoders||Chapter 5|
|4||6th November||Learning the Past Tense||Chapter 6|
Block Practical classes will be held online with Microsoft Teams.
McClelland, J. L. & Rumelhart, D. E. (1988). Explorations in Parallel Distributed Processing: A Handbook of Models, Programs and Exercises. Cambridge, MA: MIT Press.
McLeod, P., Plunkett, K., & Rolls, E. T. (1998). Introduction to Connectionist Modelling of Cognitive Processes. Oxford, UK: Oxford University Press.
Minsky, M. & Papert, S. (1969). Perceptrons. Cambridge MA: MIT Press.
O’Reilly, R. C. & Munakata, Y. (2000). Computational Explorations in Cognitive Neuroscience: Understanding the Mind by Simulating the Brain. Cambridge, MA: Bradford Books.
Plunkett, K. (2018). Handbook for Neural Network Modelling of Cognition. (see Handbook)
Rumelhart, D., Hinton, G., & Williams, R. (1986). Parallel Distributed Processing: Explorations in the Microstructure of Cognition, volume 2: Psychological and Biological Models, chapter 8: Learning Internal Representations by Error Propagation, (pp. 318–362). The MIT Press.