Course: Modern Robots - Evolutionary Robotics
T,Th 2:45-4:00pm
Room: Engineering 2070
Office hours: M 12:45-2:45, T 4-4:30,Th 12:45-2:45 and 4-4:30
Syllabus (pdf)
Homework Assignments
Below I have put the Sunday date, but technically the homeworks are due the following day (the Monday) before 5am.
If the link does not work, the homework has not been posted yet. Please check back often and let me know if you think it should have already been posted.
Homework One (pdf) (code files) Due: 1.31
Homework Two (pdf) (code files) Due: 2.07
Homework Three (pdf) (code files) Due: 2.14
Homework Four (pdf) (code files) Due: 2.21
Homework Five (pdf) (code files) Due: 2.28
Homework Six (pdf) (code files) Due: 3.06
Homework Seven (pdf) (code files) Due: 3.20
Homework Eight (pdf) (code files) Due: 3.27
Reading Assignments
To be read before the class for which they are listed. Assignments more than a few days out are subject to frequent change, so check back often. Note: If a link to the lecture slides does not work, they have not been posted yet. Please try back later.T, 1/26: No reading. First day of class. (slides)
R, 1/28: BAI 1-35 (stop after simulated annealing section), 39-42 (stop at 1.13) (slides)
T, 2/02: Biologically Inspired Computing (pdf), Why Evolutionary Robotics Will Matter (pdf) (slides)
R, 2/04: BAI 42 (1.13) to 66 (finish 1.19.1) (slides)
T, 2/09: BAI 67-97 (thru 1.25, 1.25 is optional), and Principles of modularity, regularity, and hierarchy for scalable systems (pdf) (slides)
R, 2/11: A taxonomy for artificial embryogeny. (pdf) (slides)
T, 2/16: On the performance of indirect encoding across the continuum of regularity (pdf) (slides)
R, 2/18: Compositional Pattern Producing Networks, by Stanley (pdf)(slides)
T, 2/23: A Hypercube-Based Indirect Encoding for Evolving Large-Scale Neural Networks (pdf) (slides)
R, 2/25: BAI 163–191 (stop before sec. 3.5) (slides)
T, 3/01: BAI 191–219 (stop before sec. 3.8) (slides)
R, 3/03: BAI 220–250 (stop before sec. 3.11) (slides)
T, 3/08: Catch up on reading, and play with the DNA tool at PicBreeder.com and this L-Systems tool (slides)
R, 3/10: Fitness Sharing and Niching Methods Revisited, by Sareni and Krahenbuhl 1998 (pdf) [skip/skim sections IV and V] (slides)
T, 3/15: Spring Break
R, 3/17: Spring Break
T, 3/22: Abandoning objectives: Evolution through the search for novelty alone., By Lehman and Stanley. (slides)
R, 3/24: On the Deleterious Effects of A Priori Objectives on Evolution and Representation, by Woolley & Stanley (slides)
T, 3/29: Evolving a Diversity of Virtual Creatures through Novelty Search and Local Competition, By Lehman and Stanley (slides)
R, 3/31: MAP-Elites, by Mouret & Clune (slides)
T, 4/05: Innovation Engines: Automated Creativity and Improved Stochastic Optimization via Deep Learning, by Nguyen, Yosinski, and Clune (slides)
R, 4/07: Midterm Presentations
T, 4/12: Encouraging Behavioral Diversity in Evolutionary Robotics: an Empirical Study, by J.-B. Mouret and S. Doncieux (slides)
R, 4/14: Evolutionary multi-objective optimization: a historical view of the field, by Coello Coello. (slides)
T, 4/19: Resilient Machines Through Continuous Self-Modeling, by Bongard, Zykov, and Lipson in Science (slides)
R, 4/21: The Transferability Approach: Crossing the Reality Gap in Evolutionary Robotics, by Koos, Mouret, and Doncieux (slides)
T, 4/26: Fast Damage Recovery in Robotics with the T-Resilience Algorithm, by Koos, Cully, & Mouret (slides)
R, 4/28: Robots that can adapt like animals, by Cully, Clune, Tarapore, and Mouret. Also watch the video summary. (slides)
T, 5/03: The Evolutionary Origins of Modularity, by Clune, Mouret & Lipson, and the video summary of Evolving neural networks that are both modular and regular: HyperNEAT plus the Connection Cost Technique. (slides)
R, 5/05: Natural selection fails to optimize mutation rates for long-term adaptation on rugged fitness landscapes., by Clune et al. (required for graduate students, optional for others) (slides)
T, 5/12: Final Exam (attendance required) 3:30 pm - 5:30 pm