Table Of Content

Title Page Page: iii Copyright Page Page: iv Contents Page: v Preface and Acknowledgments Page: ix 1. What Is Cognitive Modeling? Page: 1 The Use of Models Page: 1 Time Scales of Modeling Page: 5 Striving for a Goal Page: 5 Optimization Page: 8 TensorFlow Page: 13 Minimizing Energy or Getting Groceries Page: 13 2. Decision Making Page: 17 Minimization in Activation Space Page: 17 A Minimal Energy Model Page: 21 Cooperative and Competitive Interactions in Visual Word Recognition Page: 25 The Hopfield Model Page: 27 Harmony Theory Page: 29 Solving Puzzles with the Hopfield Model Page: 31 Human Memory and the Hopfield Model Page: 31 The Diffusion Model Page: 33 The Diffusion Model in Psychology Page: 35 3. Hebbian Learning Page: 37 The Hebbian Learning Rule Page: 37 Biology of the Hebbian Learning Rule Page: 40 Hebbian Learning in Matrix Notation Page: 41 Memory Storage in the Hopfield Model Page: 44 Hebbian Learning in Models of Human Memory Page: 48 4. The Delta Rule Page: 53 The Delta Rule in Two-Layer Networks Page: 53 The Geometry of the Delta Rule Page: 57 The Delta Rule in Cognitive Science Page: 61 The Rise, Fall, and Return of the Delta Rule Page: 65 5. Multilayer Networks Page: 69 Geometric Intuition of the Multilayer Model Page: 69 Generalizing the Delta Rule: Backpropagation Page: 72 Some Drawbacks of Backpropagation Page: 74 Varieties of Backpropagation Page: 76 Networks and Statistical Models Page: 82 Multilayer Networks in Cognitive Science: The Case of Semantic Cognition Page: 83 Criticisms of Neural Networks Page: 85 6. Estimating Parameters in Computational Models Page: 89 Parameter Space Exploration Page: 89 Parameter Estimation by Error Minimization Page: 91 Parameter Estimation by the Maximum Likelihood Method Page: 92 Applications Page: 99 7. Testing and Comparing Computational Models Page: 107 Model Testing Page: 108 Model Testing across Modalities Page: 114 Model Comparison Page: 116 Applications of Model Comparison Page: 120 8. Reinforcement Learning: The Gradient Ascent Approach Page: 123 Gradient Ascent Reinforcement Learning in a Two-Layer Model Page: 124 An N-Armed Bandit Page: 126 A General Algorithm Page: 127 Backpropagating RL Errors Page: 129 Three- and Four-Term RL Algorithms: Attention for Learning Page: 129 9. Reinforcement Learning: The Markov Decision Process Approach Page: 133 The MDP Formalism Page: 134 Finding an Optimal Policy Page: 138 Value Estimation Page: 138 Policy Updating Page: 142 Policy Iteration Page: 143 Exploration and Exploitation in Reinforcement Learning Page: 143 Applications Page: 145 Combining Gradient-Ascent and MDP Approaches Page: 149 Reinforcement Learning for Human Cognition? Page: 151 Open AI Gym Page: 151 10. Unsupervised Learning Page: 153 Unsupervised Hebbian Learning Page: 153 Competitive Learning Page: 156 Kohonen Learning Page: 158 Auto-Encoders Page: 161 Boltzmann Machines Page: 162 Restricted Boltzmann Machines Page: 166 11. Bayesian Models Page: 173 Bayesian Statistics Page: 173 The Rational Approach Page: 179 Bayesian Models of Cognition Page: 182 12. Interacting Organisms Page: 191 Social Decision Making Page: 191 Combining Information Page: 193 Game Theory Page: 193 Cultural Transmission and the Evolution of Languages Page: 198 To Conclude Page: 201 Conventions and Notation Page: 203 Glossary Page: 205 Hints and Solutions to Select Exercises Page: 207 References Page: 219 Index Page: 243

Description:

An introduction to computational modeling for cognitive neuroscientists, covering both foundational work and recent developments. Cognitive neuroscientists need sophisticated conceptual tools to make sense of their field’s proliferation of novel theories, methods, and data. Computational modeling is such a tool, enabling researchers to turn theories into precise formulations. This book offers a mathematically gentle and theoretically unified introduction to modeling cognitive processes. Theoretical exercises of varying degrees of difficulty throughout help readers develop their modeling skills. After a general introduction to cognitive modeling and optimization, the book covers models of decision making; supervised learning algorithms, including Hebbian learning, delta rule, and backpropagation; the statistical model analysis methods of model parameter estimation and model evaluation; the three recent cognitive modeling approaches of reinforcement learning, unsupervised learning, and Bayesian models; and models of social interaction. All mathematical concepts are introduced gradually, with no background in advanced topics required. Hints and solutions for exercises and a glossary follow the main text. All code in the book is Python, with the Spyder editor in the Anaconda environment. A GitHub repository with Python files enables readers to access the computer code used and start programming themselves. The book is suitable as an introduction to modeling cognitive processes for students across a range of disciplines and as a reference for researchers interested in a broad overview.