University of Hertfordshire School of Computer Science
Module Guide 2011-2012:
module title: Artificial Life
module code: 7COM0188
semester A - 30 credits
Instructors: Prof. Chrystopher Nehaniv (LC257, ext. 4470, C.L.Nehaniv@herts.ac.uk) - Module Leader
Prof. Kerstin Dautenhahn (LC252, ext. 4333, K.Dautenhahn@herts.ac.uk)
Dr. Joe Saunders (E114, ext. 1031, J.1.Saunders@herts.ac.uk)
Introduction to module
The overall aim of this module is to provide an in-depth study of a range of advanced ideas, theory, and techniques used in the construction of artificial life systems. The module will be oriented towards (1) the modelling of real-life biological systems and (2) the application of ideas and principles from biology and evolution to computer science in the areas of optimisation, intelligent agents, and engineering, and feedback back to the biological sciences. There is a large practical element to the module with the students gaining experience in developing artificial life models.
Module Aims:
To enable students to:
- develop skills in the building of Artificial Life systems through use of appropriate programming languages, tools and methodologies;
- investigate the application of Artificial Life (AL) techniques to optimization, to understanding biological systems, and to agent modelling (both in software and hardware);
- appreciate relevant current research topics in the theory and practice of Artificial Life;
- appreciate a range of advanced ideas and techniques modelling the properties of living systems and the exploitation of these techniques in computer science and its applications.
Module Learning Outcomes:
Knowledge and Understanding: Successful students will typically have knowledge and understanding of a variety of AL techniques and methods applicable across domains ranging from molecular computational biology and evolution of agents to behaviour-oriented and social robotics.
Skills and Attributes: Successful students will be able to critically evaluate some recent Artificial Life paradigms for building agent systems and modelling biological systems.
Semester A Times:
4 hrs lecture per week plus robotics lab which will meet an additional 2 hours / week throughout the semester.
NOTE LOCATION AND TIMES:
Lectures:
Wednesdays 10 am - 12 noon in C152
Fridays 9 am - 11 am in C402 (9-10 am) & C400 (10-11 am)
Robotics Lab: Thursdays 10 am -12 noon in E251 Mezzanine Robotics Lab
Students should plan to attend 6 hours within class each week (4 lectures + 2 hours lab) starting 5 October 2011 (Wednesday) -16 December 2011 (Friday) and also 11 January 2012 to 13 January 2012.
Coursework: In-Class Worksheets, Lab practicals, plus two projects in different areas of Artificial Life. Assessment is via coursework only, there is no exam. However the two individual projects will require substantial work! To pass this module you must pass overall. (See below for details.)
Students enrolling should have programming ability in at least one computer language.
SEMESTER A (Lectures 11 weeks in 2011 & 1 week in 2012, plus weekly robotics lab starting 13 October 2011)
Introduction (1 week): 5 & 7 October
Definitions of Life, Emergence, Bottom-up vs. Top-down approaches, Logic of Life vs. Embodied Artificial Life, Overview of Major Issues (Emergence, Self-Organization, Darwinian Evolution, Life as It Could Be, Weak vs. Strong Artificial Life, etc); Braitenberg vehicles.
Cellular Automata (1/2 week): 12 October (Wednesday)
models of natural systems (e.g. Hodge-Podge machine, Belousov-Zhabotinsky reaction), classes of CAs / computation, homeostasis, edge of chaos, examples, excitable media, topics from: synchronous and asynchronous cellular automata, lamba-parameter, dimensionality, topologies, automata networks, software simulation tools
Life and Evolution on Earth and in the Computer (1/2 week): 14 October (Friday)
Digital Organisms in Tierra. Self-Replication, Biological background for Computer Scientists. Autopoiesis.
Swarm Intelligence (1 week): 19 & 21 October
Stigmergy, optimization, sorting, collective building and maintenance, flocking, anonymous social intelligence, social interaction, anonymous vs. individualized social intelligence. Topics selected from: Robotic swarms; Swarms, Flocks, Herds, Boids; Multicellular Behaviour
* * * 24 October (Monday): Proposal for Project A Due (2-3 pages)
Face-to-Face Feedback on Proposal A within a few days (to be arranged)
Evolutionary Systems and Computation I (1/2 week): 26 October
Genetic Algorithms, Variability and Selection, Co-evolution (e.g. Hillis' sorting networks, Karl Sims's Creatures)
Robot and Agent Architectures I (1.5 weeks) : 28 October, 2 & 4 November
Introduction to behaviour-based robotics, types of robot control, reactive, subsumption, behaviour selection and modulation techniques, potential fields, and behaviour-orientation
Growth and Morphogenesis I (1/2 weeks): 9 November
L-systems (Lindenmayer systems) modelling plant and fractal growth, biological examples, environmental factors.
Evolutionary Systems and Computation II (1.5 weeks): 11, 16, & 18 November
Instances of Darwinian Evolution, Biological Examples, Genetic Algorithms, Genetic Programming (GP), Evolutionary Strategies (ES); related methods: hill-climbing and simulated annealing. Topics selected from: genetic algorithm (GA) theory (schema theorem, convergence issues, deceptive landscapes), Price equation, genotype-phenotype maps, evolvability, co-evolution, optimization of objective function vs. ecological fitness, Symbiosis, Host-Parasite Relations; Novel characters and higher levels of organization.
* * * 22 November (Tuesday): Project A due (6-8 pages in two-column IEEE format, with references (see example); + Appendices including all code
Agents & Socially Intelligent Robots I (1 week): 23 & 25 November
Definitions, embodied agents, classification, issues of autonomy and design, degrees of embodiment, social intelligence.
Growth and Morphogenesis II (1 week): November 30 and 2 December
Growth and form in nature; Approaches to morphogenesis: mathematical and physico-geometric methods (D'Arcy Thompson), L-systems, Diffusion-Reaction (Turing systems), positional information (L. Wolpert). models of multicellular morphogenesis (e.g. Marée, Glazier-Grane, etc.); What is evo-devo? Evo-devo of life on earth, Topics from: Evolution of Multicellularity, Evolution of Developmental Genetic Regulatory Networks (DGRNs) [a non-Von Neumann novel computational paradigm!]; biological complexity, duplication-divergence and division of labour; Transitions in Fitness: sex, repair and multicellular cooperation (L. Buss, R. Michod)
* * * 5 December (Monday): Proposal for Project B Due (2-3 pages) - Feed-back within same week
Self-Reproduction (1/2 week): 7 December
Self-replicators, self-reproducers, and their evolution (J. von Neumann, C. Langton, H. Sayama, and others),
Robot and Agent Architectures II (1.5 weeks): 9, 14 & 16 December
Machine learning techniques for robot learning, ontogenetic approaches to robotics, principles of developmental robotics, social learning and imitation.
* * * 15 December (Thursday): Trashcan Robotics Lab Task Due
Agents and Socially Intelligent Robots II (1 week): 11 & 13 January
Life-like believable agents, varieties of social intelligence, Human-Robot Interaction (HRI) and applications, survey of issues and example implementations. Selected advanced topics.
* * * 18 January (Wednesday): Project B Due
Important note: Coursework arrangements are subject to the approval of the external examiner and may modified. This courseplan is provisional. The dates for coursework and lectures may be subject to minor modification.
Assessment
Coursework only: Two project reports in different areas of Artificial Life, plus problem worksheets and robotics lab work
Project Report A: 35%
Project Report B: 35%
Robotics Lab: 20% (10% Weekly Effort & Participation, 10% Final Task due 15 December 2011)
In-Class Problem Worksheets: 10%
Project Reports.
For each project report, choose an area and appropriate technique(s) to do an individual project and write-up it as a report.
Break down of report marking (within each report):
The above breakdown is applied to both the report proposal (due at least several weeks before each final report) and final report (for both reports). The report proposal is worth 10% of the total report mark, while the final report is worth 90%.
Together with other coursework, the reports will be the basis of assessment and so should give evidence documenting the above, demos and appendices provide further evidence to support the report contents.
Format: Reports length will be 6-8 pages in two-column IEEE format, to which, in addition, all computer code and results must added appended. Students will be given an opportunity to demonstrate their work as further evidence of achievement.
The two reports must focus on different areas of artificial life.
Note: Either of the project reports could form the basis for extension to a full-blown MSc project.
In-class worksheets (usually about one per week, so about 10-12 in total) will be done to support students' engagement and mastery of the lecture material. They are worth about 1 mark each of the overall module grade - to get this mark the student must make an effort at the problems given.
Feedback and Learning Outcomes:
Feedback on in-class worksheets is provided during lectures in the form of solutions at the whiteboard and class discussion.
For all other coursework (proposals, reports, lab-task) written feedback provided on the marked work. In addition, oral feedback is provided on the first project proposal in a one-to-one discussion with the student.
We aim to return all coursework within one week (or in the next week of class following the hand-in date in the case of the 15 December Robotics Lab Task).
Each coursework assignment assesses all the learning outcomes.
DATES OF CLASS:
Students who attend and take notes on all lectures are most likely to succeed in this course, historically speaking.
The first day of class is Wednesday, 5 October 2011.
The last day of class in 2011 is Friday, 16 December 2011.
The first day of class in 2012 is Wednesday, 11 January 2012.
The last day of class is Friday, 13 January 2012.
Labs run on Thursdays from 13 October 2011 to 15 December 2011.
Important note: Coursework arrangements are subject to the approval of the external examiner and may modified.
This courseplan is tentative. The dates for coursework and lectures may be subject to modification.
Study time
Total: 300 hours of directed study made up of:
• class contact: 68 hours
• outside class independent study: 232 hours
Course materials and reading
Module materials and reading will be made available by the tutoring staff as HTML or PDF or hardcopy,
or pointers to required on-line material and software. There is no required course text book.
Required course readings will be handed out in lectures or on-line, and should be read prior to the next lecture.
Additionally, the students may benefit from additional reading that will be recommended in the course of the lecture, but this reading will be optional and not handed out.
Module guide moderation
I have examined this module's guide and all the sections have been completed with appropriate information. I confirm that the moderation process been completed satisfactorily. [The printed signed copy of this module guide must be lodged in the module box by the module leader.]
Signature (moderator):
Name (moderator): Dr. Neil Davey
Date: 4 October 2011