In this series on cognitive machine learning we explore the connections and mutual inspirations between cognitive science and machine learning. The is a series of four essays to complement the previous series in learning in brains and machines, but instead of taking inspiration from neuroscience, will instead use psychology and social science as its basis for inspiration. This series also explores the role that the arts can have in helping us visualise and emotionally connect to scientific concepts, and how the arts can be used in scientific communication.
- Introducing the framework that will be used to connect cognitive science and machine learning, as well as reviewing the popular frameworks for understanding cognitive systems.
- Learning to Explain
- Reasoning can be split into deduction, induction and abduction. Abduction is about providing explanations and we examine the cognitive principles of abductive inference and how explanations of events can be achieved in machine learning.
- Uncertain thoughts
- Meta-cognition, cognition about cognition, or thinking about thinking, is a core concept in cognitive science. People continuously self-evaluate their memories, decisions and attitudes, and make assessments of confidence and uncertainty. This post explores meta-cognitive processing and its potential realisations in machine learning.