Class information

Room: Science Center 252
Section 1: T, TH 1:15–2:30pm, Lab F 2:00–3:30
Section 2: T, TH 2:40–3:55pm, Lab F 3:30–5:00
Professor: Lisa Meeden
Office: Science Center 243
Phone: 328-8565

Introduction

This seminar will explore the topic of adaptive robotics with a special focus on developmental robotics, a newly emerging paradigm of research. The goal of developmental robotics is to create intelligent robots by allowing them to go through a developmental process, rather than being directly programmed to solve a particular task. By endowing a robot with an appropriate initial control architecture and adaptive mechanisms, it can learn through interactions with the world, developing self-organized knowledge about itself and its environment. We will be studying the following sorts of questions: What should be innate in the robot? What adaptive mechanisms are needed? What motivates the robot to act?

Goals for the course

• Learn how to think like a computer science researcher.
• Analyze and critically discuss research papers on adaptive robotics both in writing and in class.
• Understand the fundamental questions in the field of adaptive robotics and be able to analyze different approaches to answering these questions.
• Formulate and evaluate a research question about adaptive robotics by proposing, implementing and testing a project.
• Relate your project to prior research via a review of related literature.
• Orally present a clear and accessible summary of your project.
• Write a coherent, complete paper describing and evaluating your project. You can see examples of past projects at CS81 Spring 2010.

Grading

• Class Participation and Paper Summaries: 25%
• Labs: 35%
• Final Project: 40%
  • Proposal 5%
  • Checkpoint demonstration 5%
  • Presentation 10%
  • Paper 20%

Paper Summaries

Read all of the papers for the week prior to class on Tuesday. Bring a typed summary of the papers with you to class on Tuesday. You will turn this in at the end of class on Tuesday. I will write comments on it and return it to you at the start of class on Thursday.

For each paper include the following:

• A 1-2 paragramph overview that includes what kind of paper it is (review, assessment study, exploratory study, or a manipulation experiment) and describes its main points.
• At least three questions or issues that we can discuss during class. These might include questions of clarification, such as:
  • What is the meaning of some specific terminology?
  • How does a particular algorithm work? Why is a certain step of an algorithm necessary?
  • Clarify what is being shown in a particular figure or table in the paper.
  But should also include bigger picture issues, such as:
  • In what ways does the mechanism described seem like a good tool for doing adaptive robotics? What are its advantages/distadvantages?
  • Do the paper's conclusions logically follow from the data provided?
  • How do the paper's mechanisms and conclusions relate to previous papers that we have read?
  • What follow up experiments should be done and why?

Each Thursday we will create a summary of our discussion for that week on a wiki page: Section 1 and Section 2. The first time you go to the wiki page for your section you'll need to register. You will then be sent a password that you can use to login. Change this password by updating your profile.

---

Schedule

WEEK	DAY	ANNOUNCEMENTS	TOPIC & READING	LAB
1	Jan 17		Introduction and Neural computation Empirical Research from Empirical Methods for Artificial Intelligence, Paul Cohen (1995) pages 1-8 A Neural Network Primer from Computational Developmental Psychology, Thomas Schultz (1995) pages 1-36, pages 37-48	Lab 1: Neural Network Brains for Robots
	Jan 19
2	Jan 24		Evolutionary robotics Trends in evolutionary robotics, Chapter 9 in Soft computing for intelligent robotics systems, Lisa Meeden and Deepak Kumar (1998) Evolving 3D morphology and behavior by competition in Artificial Life IV Proceedings, Karl Sims (1994) Automatic design and manufacture of robotic lifeforms, Nature, Vol. 406, Hod Lipson and Jordan Pollack (2000)	Lab 2: Scribbler robot
	Jan 26	Drop/Add ends (Jan 27)
3	Jan 31		NEAT and Novelty search Competitive coevolution through evolutionary complexification Journal of Artificial Intelligence Research, Vol. 21, Kenneth Stanley (2004) Abandoning objectives: Evolution through the search for novelty alone Evolutionary Computation, Vol. 19:2, Joel Lehman and Kenneth Stanley (2011)	Lab 3: Evolving with NEAT
	Feb 02
4	Feb 07		Developmental robotics Introduction to developmental robotics Connection Science, Vol. 18.2, Lisa Meeden and Douglas Blank (2006) Some basic principles of developmental robotics IEEE Transactions on Autonomous Mental Development, Vol. 1.2, Alexander Stoytchev (2009) Developmental robotics: A survey Connection Science, Vol. 15.4, Max Lungarella, Giorgio Metta, Rolf Pfeifer, and Giulio Sandini (2003)	Lab 4: Novelty search
	Feb 09
5	Feb 14		Segmentation and Categorization Continuous categories for a mobile robot Proceedings of AAAI, M. Rosenstein and P. Cohen (1999) Sensory flow segmentation using a RAVQ Advances in Pattern Recognition, Fredrik Linaker and Lars Niklasson (2000) A behavior-based approach to forming categories IEEE Transactions on Autonomous Mental Development, S. Griffith, J. Sinapov, and A. Stoytchev (2011)	Lab 5: Unsupervised categorization
	Feb 16
6	Feb 21		Intrinsic motivation Intrinsic motivation systems for autonomous mental development IEEE Transactions on Evolutionary Computation, vol. 11.2, Peirre-Yves Oudeyer, Frederic Kaplan, Verena Hafner (2007) Category-based intrinsic motivation, Proceedings of the Ninth International Conference on Epigenetic Robotics, Rachel Lee, Ryan Walker, Lisa Meeden, and James Marshall (2009) A growing neural gas learns topologies Advances in Neural Information Processing Systems 7, Bernd Fritzke (1995)	Lab 6: Intrinsic motivation
	Feb 23
7	Feb 28		Reinforcement learning Developing navagation behavior through self-organizing distinctive state abstraction by Jefferon Provost, Benjamin Kuipers, and Risto Mikkulainen, Connection Science, Vol. 18.2, 2006	None
	Mar 01
	Mar 06	Spring Break
	Mar 08
8	Mar 13		Grounding language TED talk by Deb Roy: The birth of a word Grounded situation models for robots: Where words and percepts meet by Nikolaos Mavridis and Deb Roy, Intelligent Robots and Systems, 2006 A habit system for an interactive robot by Kai-yuh Hsiao and Deb Roy, AAAI Fall Symposium, 2005 Semiotic schemas: A framework for grounding language in action and preception by Deb Roy, Artificial Intelligence, Vol. 167:1-2, 2005	Lab 7: Controlling physical robots
	Mar 15
9	Mar 20		Social interaction Robots that imitate humans, by Cynthia Breazeal and Brian Scassellati, Trends in Cognitive Sciences, Vol. 6:11, 2002 Early integration of vision and manipulation by Giorgio Metta and Paul Fitzpatrick, Adaptive Behavior, Vol. 11:2, 2002	Final project
	Mar 22	Last day to declare CR/NC or W (Mar 23)
10	Mar 27		Group behavior Emergence of communication in embodied agents evolved for the ability to solve a collective navigation problem, by Davide Marocco and Stefano Nolfi, Connection Science, Vol 19:1, 2007 Towards group transport by swarms of robots, by Roderich Gross and Marco Dorigo, International Journal of Bio-Inspired Computation, Vol. 1:1-2, 2009	Discuss project proposals
	Mar 29
11	Apr 03		Implementing imitation Hierarchical attentive multiple models for execution and recognition of actions, by Yiannis Demiris and Bassam Khadhouri, Robotics and Autonomous Systems, Vol. 54, 2006	Work on final project
	Apr 05
12	Apr 10		Checkpoint Demonstrations
	Apr 12
13	Apr 17		Work on final project
	Apr 19
14	Apr 24		Presentations section 1 Adrien, Danny, Sam Catie, Stella Mark, Steve Presentations section 2 Temple Celia, Lisa Dasol, Emily	Presentations section 1 David W., John Kristen, Peter Presentations section 2 Dylan, Greg, Jessi Alex B., Seth
	Apr 26		Presentations section 1 Taylor Abigail, Bernie David D., Kenny Presentations section 2 Abel, Alex C. Emanuel Nick, Phil

Final Projects

• Music genre classification using machine learning techniques, Sam Clark, Danny Park, and Adrien Guerard
• Hierarchical attentive multiple models for execution and recognition of actions through training inverse and forward models explorationally, Alex Cannon and Imoleayo Abel
• Neato Quadcopters, Alex Burka and Seth Foster
• Learning from experience: Response of Q-learning to the saliency of environmental cues, Catie Meador and Stella Cho
• Grounding color and shape words in visual context, Celia Abernathy and Lisa Bao
• NOVELEAT: Combining novelty search and objective-based search, John Dinh and David Wilokofsky
• Learning classes of melodies with a recurrent neural network, David D'Annunzio and Kenneth Ning
• Evolving combat robots, Dylan Katz, Greg Rawson, and Jessi Velasco
• Categorizing the environment for improved performance at a light eating task, Emanuel Schorsch
• Evolving grounded communication in a navigation task using NEAT, Kristen Allen and Peter Ballen
• Combining Q-learning with artificial neural networks in an adaptive light seeking robot, Steve Dini and Mark Serrano
• Evolving flocking navigation controllers for adaptive swarms, Nicolas Felt and Phillip Koonce
• Categorization of dynamic environments with CBIM, Taylor Nation
• Methods of communication, Gene T. Price
• Competitive coevolution of robots to play ultimate frisbee, Abigail Henderson and Bernie Koch
• Applying neural pruning to NEAT, Emily Dolson and Dasol Park