Exploring Computational Theory of Mind

Sep 12, 2024

Lecture Notes: Computational Theory of Mind and Brain

Introduction

  • Topic: Mars Computational Theory Level of Analysis
  • Focus: Color vision as a case study
  • Agenda: 1) Discuss color vision in terms of computational theory 2) Methods in cognitive neuroscience, focusing on face perception

Key Themes

  • Big Question: How does the brain give rise to the mind?
  • Physical Basis of the Brain: Previous lectures addressed the brain's physical structure.
  • Defining the Mind:
    • The mind is viewed as a set of computations that extract representations.
    • Mental representations can be perceptual (e.g., seeing motion) or conceptual (e.g., thoughts).

David Marr's Framework

  • Marr's Big Idea: Understand what is computed and why at the level of computational theory.
  • Example of Vision:
    • Light enters the eyes, leading to perception.
    • Goal: Understand how computations convert light (input) to perception (output).

Ill-posed Problems in Perception

  • Color Vision Problem:
    • Determining an object's color (reflectance) from the light it reflects (luminance) can be ill-posed.
    • Solution involves making assumptions about the light illuminating the object.
  • Other Examples:
    • Shape perception is also an ill-posed problem: multiple shapes can cause the same retinal image.
    • Language and meaning can be ill-posed (e.g., ambiguous words).

Computational Theory in Color Vision

  • Using Color: Color helps in identifying objects (e.g., fruit) and their properties.
  • Ill-posed Nature: We often lack enough information to deduce color directly; need additional context.
  • Assumptions and Knowledge: Necessary to deduce color from luminance.

Levels of Analysis

  1. Computational Theory: What is computed and why?
  2. Algorithm and Representation: What is the code to solve the computational problem?
  3. Hardware: Neural basis of the computations (e.g., neurons involved in color perception).

Psychophysics and Behavioral Studies

  • Psychophysics: Assessing how people perceive different stimuli and their properties.
  • Demo on Color Recognition: Different colored cars perceived as different colors despite being the same color due to context cues.

Face Perception Case Study

  • Importance of Face Recognition:
    • Vital for social interactions; conveys identity, mood, and other personality traits.
    • Example: Jacob's experience with prosopagnosia (face blindness).
  • Questions in Face Recognition:
    • How does face recognition differ from object recognition?
    • What computational processes are involved?

Tools for Face Recognition Research

  1. Behavioral Data: Reveal how we recognize faces based on experience and familiarity.
  2. Functional MRI:
    • Measures brain activity based on blood flow changes (BOLD signal).
    • High spatial but low temporal resolution.
    • Useful for comparing brain responses to faces versus objects.

Conclusion

  • Understanding the computational theory of mind requires multiple levels of analysis and diverse methods.
  • Face perception remains a rich field of study with ongoing research to understand the neural basis and computational mechanisms involved.

Note: Further readings and assignments on functional MRI and cognitive neuroscience methods will be provided next class.