Artificial Cognitive Systems

Artificial Cognitive Systems - A Primer, MIT Press, 2014


Course Description  |  Learning Objectives  |  Outcomes  |  Content  |  Lecture Notes  |  Course Textbook  |  Recommended Reading  |  Useful Links

Course Description

The goal of this course is to expose students to a comprehensive cross-section of the main elements of artificial cognitive systems, a discipline which draws on artificial intelligence, developmental psychology, and cognitive neuroscience. Concepts are introduced in an intuitive, natural order, with an emphasis on the relationships among ideas and building to an overview of the field, equipping students with sufficient knowledge and understanding to study specific topics in greater depth. The course is delivered through a mix of teaching, reading, and in-class discussion. A particular feature of the course is its emphasis on progressive deepening: covering topics several times in increasing detail as the course proceeds. This approach helps the student develop her or his understanding each topic, and the relation of that topic to other topics. Student progress is assessed by a series of multiple choice tests and written individual & group assignments. There are no prerequisites for this course.

Student Elevator Pitch for the Course

To find out what students who took the course thought of it, in 2019 I ran a competition at Carnegie Mellon University Africa in which each student made a pitch to hypothetical prospective students. The competition was judged by the other students and it was won by Timothy Odonga. Here is his pitch.

"There are three types of students:
Student Type A: You have taken computer vision, machine learning, or deep learning and you are wondering what to do with all this knowledge.

Student Type B: You are interested in AI, but you don't know where to start. What do deep learning, machine learning, and computer vision achieve in the context of AI? What is a good place to start?

Student Type C: You're not one of the above two types but you are looking for a class outside your normal course profile.

This is what each student will gain from the course:
Student Type A: ACS brings everything together. The material you have learned will make more sense in the broader picture. The goal of AI is to build intelligent systems.

Student Type B: You will find more focus into what field you find interesting and want focus on in AI from cognitive robotics, deep learning, machine learning, and neuroscience.

Student Type C: You will learn a lot and build on your general knowledge. You will learn a lot of big words, and gain exposure to psychology, epistemology, nonlinear dynamics, and bit of information theory.

All three types of students will learn skills such as verbal and nonverbal communication, presentation skills, how to critique a research paper, and handling interviews.

The class focuses on the principles in building cognitive systems and incorporate the necessary cognitive capabilities, understanding the cognitive paradigms, and cognitive architectures. There is a progressive deepening of concepts presented that makes the learning a lot easier."

Learning Objectives

Students will learn about the nature of cognition and the motivation for studying artificial cognitive systems and they will be introduced to different ways to modelling cognitive systems. Students will learn about the three paradigms of cognitive science: the cognitivist paradigm, the emergent paradigm, and the hybrid paradigm. They will examine cognitive architectures in some detail, learning about the role of a cognitive architecture, its desirable characteristics, its core cognitive abilities, and the different options when designing one. They will study several representative cognitive architectures at various levels of detail. Students will learn about autonomy, both robotic and biological, and they will be introduced to the concepts of constitutive autonomy, behavioural autonomy, homeostasis, allostasis, and self-organization. They will learn about embodiment and the various hypotheses on embodiment, including empirical evidence to support these hypotheses. They will study development and learning and how these two processes differ, including motivation, drives, and value systems. Students will learn about the different forms of memory and how these are involved in different types of anticipation, through self-projection, prospection, and internal simulation. They will learn about the different stances taken on knowledge and representation, and the symbol grounding problem. Finally, they will learn about social cognition, social interaction, and the relevance of joint action, shared goals, shared intention, and joint attention. In this context, they will also learn about reading intentions, theory of mind, instrumental helping, collaboration, interaction dynamics, and different schools of thought on cognitive development.


After completing this course, students should be able to:

  1. Identify the key attributes of a cognitive system.

  2. Explain the main characteristics of cognitivist, emergent, and hybrid cognitive science.

  3. Compare cognitive architectures using several criteria and design an outline cognitive architecture for a given application scenario.

  4. Explain how a specific hybrid cognitive architecture works and show how it can be used to allow a robot to reason about its environment and achieve goals set by a user.

  5. Explain the implications of computational functionalism and its relationship to the embodied cognition thesis.

  6. Distinguish between learning and development and explain how these processes are facilitated by different forms of memory and knowledge.

  7. Distinguish between instrumental helping, cooperation, and collaboration.

  8. Explain how social cognition relies on the ability to take perspectives and form a theory of mind.

  9. Explain the interplay between of joint action, shared goals, shared intention, and joint attention in social cognition.

Lecture Notes

Module 1: The Nature of Cognition
Lecture 1: Course overview. Motivation for studying artificial cognitive systems.  
Lecture 2: What is cognition? Modelling cognitive systems. Levels of abstraction and the ultimate-proximate distinction. 

Module 2: Paradigms of Cognitive Science
Lecture 1: The different paradigms of cognitive science. The cognitivist paradigm of cognitive science.  
Lecture 2: The emergent paradigm of cognitive science; connectionist, dynamical, and enactive approaches. 
Lecture 3: Hybrid approaches. A comparison on cognitivist and emergent approaches. 

Module 3: Cognitive Architecture
Lecture 1: The role of a cognitive architecture. Desirable characteristics. Core cognitive abilities. How to design a cognitive architecture.  
Lecture 2: Example cognitive architectures: Soar, ACT/R, Global Workspace, LIDA, BBD/Darwin. 
Lecture 3: Example cognitive architectures: Clarion, ISAC. 
Lecture 4: Example cognitive architectures: CRAM. 
Lecture 5: Example cognitive architectures: The Common Model of Cognition. 

Module 4: Autonomy
Lecture 1: Robotic Autonomy. Strength vs. degree. Biological autonomy. Constitutive & Behavioural autonomy. Homeostasis. Allostasis. Self-organization. 
Lecture 2: Autonomy 3: Autonomic Systems. Scales of Autonomy. Goals. Measuring Autonomy. Autonomy and Cognition. 

Module 5: Embodiment
Lecture 1: Cognitivist and emergent perspectives on embodiment. The embodied cognition thesis. Three hypotheses on embodiment.  
Lecture 2: Evidence for the embodied stance. Off-line embodied cognition. From situation cognition to distributed cognition. 

Module 6: Development and Learning
Lecture 1: Development, motivation, and imitation. Value systems. 
Lecture 2: Development vs. learning. Phylogeny vs. Ontogeny.  
Lecture 3: Development from the perspective of psychology. 

Module 7: Memory and Prospection
Lecture 1: Types of memory. The role of memory.  
Lecture 2: Self-projection, prospection, and internal simulation. 
Lecture 3: Internal simulation and action. Forgetting. 

Module 8: Knowledge & Representation
Lecture 1: The duality of memory and knowledge. Representation and anti-representation. The symbol grounding problem. 
Lecture 2: Joint perceptuo-motor representations. 
Lecture 3: Acquiring and sharing knowledge. 

Module 9: Social Cognition
Lecture 1: Social interaction. Action, goals, intention, attention. 
Lecture 2: Reading intentions and theory of mind. Instrumental helping. Collaboration. 
Lecture 3: Development and interaction dynamics. Piaget and Vygotsky. Zone of proximal development. 

Course Textbook

D. Vernon, Artificial Cognitive Systems, MIT Press (2014).

Recommended Reading

Kelly, J. E., Computing, cognition and the future of knowing, IBM Corp. 2015.

Lieto, A., Bhatt, M., Oltramari, A., and Vernon, D., "The Role of Cognitive Architectures in General Artificial Intelligence", editorial for a special issue on "Cognitive Architectures for Artificial Minds", Cognitive Systems Research, Vol. 48, pp. 1-3, 2017

Rosenbloom, P., Laird, J., and Lebiere, C. "Précis of 'a standard model of the mind'", Advances in Cognitive Systems, 5:1-4, 2017.

Vernon, D. "Cognitive Architectures", in Cognitive Robotics Handbook, A. Cangelosi and M. Asada (Eds.), MIT Press, 2022.

Vernon, D. "Two Ways (Not) To Design a Cognitive Architecture", Proceedings of EUCognition 2016, Cognitive Robot Architectures, European Society for Cognitive Systems, Vienna, 8-9 December, 2016, R. Chrisley. V. C. Müller, Y. Sandamirskaya. M. Vincze (eds.), CEUR-WS Vol-1855, ISSN 1613-0073, pp. 42-43, 2017.

Vernon, D., "Cognitive System", in Computer Vision: A Reference Guide, K. Ikeuchi (Ed.), Springer, 2014.

Vernon, D., Metta. G., and Sandini, G. A Survey of Artificial Cognitive Systems: Implications for the Autonomous Development of Mental Capabilities in Computational Agents, IEEE Transactions on Evolutionary Computation, special issue on Autonomous Mental Development, Vol. 11, No. 2, pp. 151-180, 2007.

Vernon, D. and Vincze, M. "Industrial Priorities for Cognitive Robotics", Proceedings of EUCognition 2016, Cognitive Robot Architectures, European Society for Cognitive Systems, Vienna, 8-9 December, 2016, R. Chrisley. V. C. Müller, Y. Sandamirskaya. M. Vincze (eds.), CEUR-WS Vol-1855, ISSN 1613-0073, pp. 6-9, 2017. Vernon, D., von Hofsten, C., and Fadiga, L. Desiderata for Developmental Cognitive Architectures", Biologically Inspired Cognitive Architectures, Vol. 18, pp. 116-127, 2016.

Additional Reading

Kotseruba, I. and Tsotsos, J. "40 years of cognitive architectures: core cognitive abilities and practical applications", Artificial Intelligence Review, 53(1):17-94, 2020.

Laird, J. E., Lebiere, C., & Rosenbloom, P. S. "A standard model of the mind: Toward a common computational framework across artificial intelligence, cognitive science, neuroscience, and robotics", AI Magazine, 38(4), 13-26 , 2017.

Vernon, D. "Cognitive Architectures", in Cognitive Robotics, A. Cangelosi and M. Asada (Eds.), MIT Press, Chapter 10, 2022.

Vernon, D. "Reconciling Constitutive and Behavioural Autonomy: The Challenge of Modelling Development in Enactive Cognition", Intellectica, Vol. 65, pp. 63-79. 2016.

Vernon, D., von Hofsten, C., and Fadiga, L., A Roadmap for Cognitive Development in Humanoid Robots, Cognitive Systems Monographs (COSMOS), Springer, ISBN 978-3-642-16903-8 (2010).

Useful Links

Please refer to my Wiki for links to related resources and support material, including tutorials, research networks, and degree programmes in artificial cognitive systems.

David Vernon's Personal Website