Evolutionary Robotics: Theory and Practice

Posgrado en Ciencia e Ingeniería de la Computación

Semester 01-2016, Aug. – Dec. 2015

Plan: Maestría en Ciencia e Ingeniería de la Computación
Actividad académica: Temas Selectos de Inteligencia Artificial
Tema: Sistemas multiagente reactivos agentes basados en el comportamiento
Carácter: Optativa

Horarios: Lunes y Miércoles, 11:30 – 13:00 (Primera clase: 10 de agosto 2015)
Lugar: Anexo del IIMAS, Circuito Escolar, Ciudad Universitaria, DF

You can download the official program of activities for this course.

For more information, contact Dr. Tom Froese via e-mail at t.froese AT unam.mx


Primary course material:

All of these books are available online (at least from inside UNAM) and/or as hardcopies in UNAM libraries.

Nolfi, S., & Floreano, D. (2000). Evolutionary Robotics: The Biology, Intelligence, and Technology of Self-Organizing Machines. Cambridge, MA: MIT Press

Rohde, M. (2010). Enaction, Embodiment, Evolutionary Robotics: Simulation Models for a Post-Cognitivist Science of Mind. Amsterdam: Atlantis Press

van Gelder, T., & Port, R. F. (1995). It’s about time: An overview of the dynamical approach to cognition. In R. F. Port & T. van Gelder (Eds.), Mind as Motion: Explorations in the Dynamics of Cognition (pp. 1-43). Cambridge, MA: MIT Press

Vargas, P., Di Paolo, E. A., Harvey, I., & Husbands, P. (Eds.). (2014). The Horizons of Evolutionary Robotics. Cambridge, MA: MIT Press

Secondary course material:

The books are available as hardcopies in UNAM libraries.

Harvey, I., Di Paolo, E. A., Wood, R., Quinn, M., & Tuci, E. A. (2005). Evolutionary robotics: A new scientific tool for studying cognition. Artificial Life, 11(1-2), 79-98

Port, R. F., & van Gelder, T. (Eds.). (1995). Mind as Motion: Explorations in the Dynamics of Cognition. Cambridge, MA: MIT Press.

Pfeifer, R., & Bongard, J. C. (2007). How the Body Shapes the Way We Think: A New View of Intelligence. Cambridge, MA: MIT Press

More inspiration can be found here: https://minimalcognition.wordpress.com


Software packages:

The practical part of the course will be based on software packages developed by Randall Beer at Indiana University.

Evolutionary Agents: C++ software to support the evolution of model agents controlled by continuous-time recurrent neural networks.

Dynamica: A Mathematica package for the analysis of smooth dynamical systems. Dynamica includes tools for computing and displaying the vector fields, trajectories, flows, phase portraits, bifurcation diagrams and parameter charts of dynamical systems described by sets of ordinary differential equations.

Both can be downloaded from Randall Beer’s homepage.


Lectures:

Class 1a:

Download lecture slides: EvoRob_Lecture_01

Homework: read Preface and Chapter 1 of both Nolfi & Floreano and Vargas et al.

Class 1b:

Task: install Evolutionary Agents and get the walker example running

Homework: play with the code to get fastest forward velocity and, as an extra challenge, try to make the agent move backwards

Class 2a:

Homework for next Monday: read Chapter 2 of Nolfi and Floreano

Class 2b:

Get the CTRNN to evolve as a controller of leg movement.

Class 3a:

Homework: read Chapter 3 of Nolfi and Floreano

Class 3b:

Get the CTRNN to evolve as a controller of leg movement.

Class 4a:

Homework: read Chapter 5 of Nolfi and Floreano

Class 4b:

Make the evolutionary algorithm working from a seed individual.

Class 5a/5b:

Homework: read Chapter 6 of Nolfi and Floreano and load 3-node CTRNN into Mathematica

Class 6:

Homework: read Beer’s chapter from Vargas et al. (Eds.) (2014)

Class 7a/7b:

Homework: read introduction chapter of Mind as Motion and complete Mathematica assignment

Class 8a/8b:

Homework: evolve two light emitting agents to coordinate behavior

Class 9a/9b:

No classes this week. Use this time to keep working on the programming assignments and to prepare for student presentations. Please read Chapter 7 of Nolfi and Floreano for Jorge’s preparation and Froese and Di Paolo (2010) for Leticia’s presentation.

Class 10a/10b:

Finish evolving coordinated agent behavior for practical class next Monday. Read article by Egbert and Barandiaran in preparation for class on Wednesday.

Class 11a/11b:

No classes on Nov 2 and 4.

Class 12a/12b:

Wednesday, Nov. 11: Evaluation of personal project advances
Friday, Nov. 13, 11:30am in the computer science department: Student presentations by Jorge and Leticia. Please read Williams et al. (2008) and Froese and Di Paolo (2011) in preparation for class discussion.