Learning and AI

INSTRUCTOR

Daniel Rothschild

ABOUT

The remarkable recent advances in AI are powered by learning algorithms that allow computers to develop abilities through training. Unlike traditional programming, where software is explicitly coded with specific instructions, modern AI systems learn through experience, adapting and improving their performance as they process vast amounts of data. This module uses machine learning as a window into the more familiar, but still largely mysterious, process of human learning. We will read and discuss classical and contemporary literature from a variety of disciplines including philosophy, computer science, psychology, and evolutionary biology on learning and cognition.

MECHANICS

The seminar meets Thursdays from 2pm to 5pm in the seminar room of the UCL Philosophy Department, 19 Gordon Square. There is no meeting on 30 May. On 6 June the seminar will meet from 11-1pm in order not to clash with the Shearman Lectues (Nico Kolodny) and the last session will be 13 June.

The module will be assessed by a 3500-word essay.

All unlinked readings available here (ask instructor for password).

All staff and students welcome at any sessions.

SCHEDULE

25 APRIL: LEARNING IN THEORY AND PRACTICE

2 MAY: MACHINE LEARNING AND DEEP LEARNING

9 MAY: DO HUMANS LEARN FASTER THAN MACHINES?

16 MAY: LEARNING AND MOTIVATION

23 MAY: THE STARTING STATE AND BEYOND

6 JUNE 11-1pm LEARNING AS SOCIAL

June 13 ANALOGY AND ABSTRACTION


Image: Lace pattern woodcut by Isabella Catanea Parasole, 1600