Last updated: 27 April 2012
Lecture notes for CM30229
& Cognitive Systems 2012
Lecture 1: Course Introduction,
Intelligence Broadly, & Sensing
Note: You should have free access to the papers linked below
so long as you are on campus or tunneled to campus (e.g. VPN.)
- Other Notes
Lecture 2: An Introduction to
Artificial Intelligence (its history) & Cognition
Note: these slides were modified with notes to myself from the
blackboard for next year. I fixed some spellings on those notes.
Coursework 1: Wall Following
Lecture 3: Action Selection
- Some of the lecture notes came out of a 2004 paper of mine
Joanna J. Bryson, Action
Selection and Individuation in Agent Based Modelling, in
Proceedings of Agent 2003: Challenges of Social Simulation,
David L. Sallach and Charles Macal eds., pp. 317-330, Argonne
National Laboratory, 2004. (Don't worry, I won't have my
own papers on the exam! I'd read the Brooks one before
- Other People's Notes
- A bunch of the slides for this lecture were taken from Jim Blythe's
course, which includes a lot of suggested
readings. Note that is pretty old (2002, but with
updates maybe indicating 2004) but it's still fine for the
history of the approach.
Lecture 4: Cognitive Architectures
- Derbinsky, N., Laird, J.E.: Extending Soar with Dissociated Symbolic
Memories. Symposium on Human Memory for Artificial
Agents, AISB (2010)
- This is an example of a postgraduate student paper such as
CM50230 students should be able to write for their fourth
coursework, and all the courseworks are vaguely heading
- It is also an example of one step in the progress of a
group working over decades to achieve human-level AI.
- Further, it talks about memory, which will be the topic of
next week's lecture(s).
- Anderson, J.R., Fincham, J. M., Qin, Y., & Stocco, A.
(2008). A Central circuit of the mind. Trends in
Cognitive Science. 12(4), 136-143.  [info]
- This is a paper arguing ACT-R is a model of the brain from
the highest impact journal in Cognitive Science (higher
impact than any
journal in AI.)
- Architecture home pages:
- Shanahan has a book
about his theory of consciousness (will probably move this to
a later lecture, but Shanahan got mentioned today under
- Despite being very, very influential. Subsumption
Architecture does not have a web page. See the Brooks
paper above, but the original formulation (and the pictures
from today's lecture was A
robust layered control system for a mobile robot (cited
6680 times as of Feb 2011).
Lecture 5: Perception
- Optional Reading
- Tom Mitchell's book Machine
Learning has a lot of information on line.
- If you are interested in natural intelligence like the vision
examples I had in lecture today, you might want to read
Michael Mann's book which is also available entirely on line,
System in Action. The Carlson textbook cited in
the slides is more up to date but less conveniently actually
in the library.
- Clarification on the nature of science writing: There
are two things a scientist has to do:
If you only learn & understand without communicating then
you may be living the life of the mind, but you are not doing
science because science is a social process involving the
advancement of human knowledge generally. But you will never be
able to communicate everything you have learned, so you have to
selectively work on communicating well what is most likely to be
of use to others.
- Learn & understand.
- Communicate new learning to other people.
Lecture 6: Learning
Lecture 7: Evolution and Genetic
Lecture 8: Simulation &
- Reading (for CW2)
- Bryson, J. J., Ando, Y., and Lehmann, H. (2007). Agent-based models as
scientiﬁc methodology: A
case study analysing primate social behaviour.
Philosophical Transactions of the Royal Society, B
— Biology, 362(1485):1685–1698. (just the first two sections
- Horswill, I. D. (1993). Specialization
of Perceptual Processes. PhD thesis, MIT, Department of
Cambridge, MA. (just the details if you want them,
lectures were probably enough)
- King, G. (1995). Replication,
replication. PS: Political Science and Politics,
- The Mitri paper cannot be distributed from my web page for
copyright reasons, but from campus or through the library
e-journals site you should be able to download
it free of charge here.
Coursework 2: Social Behaviour or Robot
Lecture 9: Science, Agents and
- Craig Reynolds is vaguely trying to maintain a Boids Web Page with
a list of all the applications (and "rip offs" :-). Here is the
first, original movie that Boids was used for, Stanley and Stella.
Lecture 10: Social Simulation and
Lecture 11: Hypothesis Testing
Lecture 12: Multiple Conflicting Goals
– Intro to Game AI
Coursework 3: Game AI
- HTML Version
- PDF Version
- Latex Version
- Bibtex for Latex Version
Note: for some reason the current version is hard to find with
Google, but this is the right link.
- Joanna J. Bryson ``Behavior-Oriented
Modular Agent Intelligence'' (HTML, there is also a pdf
version), Agent Technologies, Infrastructures, Tools,
and Applications for e-Services, R. Kowalszyk, J. P.
Müller, H. Tianfield and R. Unland, eds., pp. 61-76,
- or just see the BOD web
- Oh yes, and then you'll want some files...
Coursework 4: The Workshop Paper
only – undergraduates get an exam. The assignment is
straight-forward. You should extend one of the first three
courseworks to be a conference paper. Please come talk to me about
what project you want to do, and if you like I will recommend a real
conference for you to target. But the paper should be about 4-6
pages of double-column length such as is used by IJCAI,
ACM or IEEE for all of their conferences. The deadline is 6 May
To work to distinction, you should actually find some recent
workshop, conference or even journal articles and use these to
establish the current state of the art / knowledge boundary, and
then see if you can replicate and / or extend one or more of them.
Lecture 13: Culture, Language &
Cognition (double-length two-day lecture)
Lecture 14: Chatbots, Turing
Tests & Believability
Lecture 15: Emotions, Drives
& Complex Control
Lecture 16: Consciousness &
- Slides (pdf)
- Note (in case you missed the lecture) the slides from my own
paper at the end are not
examinable, they were just to let you know what my biases are.
You can also find that on my Robot
- Note to self: This s/b more about Theory of Mind &
come after Believability
Lecture 17: Ethics &