Last updated: 22 March 2018

Home Page for CM50230 & CM30229
Intelligent Control & Cognitive Systems   2018

Dr. Joanna Bryson

This unit reflects a rapidly developing field.  Therefore the schedule on this page is always subject to change.

This course has three different but complementary objectives (see unit descriptions):
  1. To give students skills and experience in constructing intelligent systems.
  2. To provide an academic survey concerning the roles and capabilities of cognition in both nature and artefacts.  The goal of this survey is that students should become conversant in the topics, be able to research and provide advice for companies, organisations, or citizens that ask for their help or opinion, and able to seek out, acquire, and/or engage with expertise on these topics.
  3. To give students skills and experience in reading and writing academic reports and communication.  This should be of benefit for writing dissertations.  For CM50230, the level of skill in writing should extend to a capacity for submission for publication in workshops or conferences.

Because this course is about a topic that is both central to human identity, and also disrupting the global economy, much of its content is necessarily a matter of controversy.  For this reason the material is presented with a strong component of history of science. This is so you as students can understand and evaluate the biases not only of the lecturer, but also of other academics you may read.  Bias is an intrinsic part of intelligence (as you'll learn in this course), the question is how to best cope with it.  This course is consequently also about how academics and societies arrive at a set of material which we present to students and use in policy.  Societies and universities are also cognitive systems, so this course is about itself, and your degree.

This unit is taught flipped classroom. This is partially in response to student reviews from the first five years, where students wanted more time to go over the lectures and to ask questions.  It is also in response to the fact that the lecturer lives on a different continent. Most weeks there are two lectures to watch, generally on video, plus one quiz and one live discussion session. The time for the quiz & the discussion session add up to one additional hour.  Some weeks there is only one video and one live tutorial for the coursework, or one or two live lectures where the video needs to be updated.

To make sure that everyone is keeping up with the videos, there are quizzes. These are on moodle but at a fixed time, generally at the beginning of the lecture discussion session.  The quizzes do not count for much of your mark, and you will drop your worst two (of seven) of your marks.  But they substantially improved performance on the final exam last year, the first year they were introduced.

This page is provided as a resource, mostly so you can find lecture notes and problem sets.  There is also a moodle page associated with this course, which contains forums for support & is where you submit your coursework.  You are responsible for reading thoroughly all communication from email, the moodle news forum, and the coursework specifications.

Note on Labs:  Labs start in the first week, but not before the first lecture; those with the lab before the first lecture should attend another lab first week.  In general you are welcome to attend labs other than your assigned one, but the people who are assigned to a lab have first precedence for equipment and tutor time.
Note on Quizzes:  Quizzes start in the second week, you will be marked on your best 5 of 7. They are (mostly) held at the beginning of lecture, those not attending a lecture cannot take that week's quiz.  Because of logistics, two quizzes will be conducted without this restriction.

Week
 Lecture Topics
(exact organization still subject to change)
Coursework & Labs
(courseworks due Friday at midnight, handed out Tuesdays in Lecture)
First Lecture
Special notes
1
Introduction, Intelligence & Sensing; Artificial Intelligence & Cognition.
Lecture 1; Lecture 2
Coursework 1 handed out. 5 Feb Don't miss the first lecture! Don't miss a first lab!  Robot kits handed out in the labs AFTER the first lecture, and we help you with the software to program them there.
No quiz this week.
2
Action Selection; Cognitive Architectures.
Lecture 3; Lecture 4
tuesday quiz: 1-3
12 Feb
This is a bunch of history, but it's very relevant to your first lab as you think about the achitecting your robots' mind.
3
Perception, Learning, and Evolution.
Lecture 5; Lecture 6
tuesday quiz: 4-6
19 Feb
We won't use learning much in the labs, but we need to understand how it's used in Nature, and how it should inform our Design.
4
Evolvability, Learnability, and Design–Perception for Robots; Science, Agents and Spatial Simulations
Lecture 7; Lecture 8
Coursework 1 Due
Coursework 2  & Coursework 4 handed out
tuesday quiz: 7, 8
26 Feb
More advanced tips from Nature on learning to learn; then moving on to thinking about cognition as a social process.
Quiz is just on learning, evolution, and design, not on agents or science.
5
Social Simulation and Social Structure. NetLogo
Lecture 9; Lecture 10;
NetLogo lecture live Tuesday, 6:15, by Joanna
Netlogo labs this week & next.


5 Mar
How AI can be used to understand intelligence in nature; then moving on to thinking about cognition as a social process.
No quiz this week.
6
Hypothesis Testing and Evidence; 
Multiple Conflicting Goals: Intro to Game AI 
Lecture 11; Lecture 12
tuesday quiz 10-12
12 Mar
This week makes sure you have what you need to write up your CW2, and then introduces the concepts for CW3. Quiz is on science and social simulations, not game AI and hypothesis testing, but will discuss this week's lectures in flipped classroom Tuesday, especially multiple conflicting goals because that leads into ethics.
7
Likeability, Believability & Engagement;  NewGame
Lecture 15; Lecture 16
Game AI lecture Tuesday 6:15pm by Andreas
Game AI labs this week & after break
Coursework 2 due TUESDAY
Coursework 3 handed out

19 Mar
The lectures for Game AI also bridge you into human-like behaviour, which will be core for thinking about ethics. Who do you put first, yourself or your team's goals? There's no simple answer.
No quiz this week. (But watch the Likeability lecture some time, it's really useful for understanding commercial game AI, may crop up in 17 April quiz.)


Break

8
Culture, Language & Cognition: I & II
Lecture 17 & 18 (double lecture)

Friday quiz 16-18  (mostly 17, 18)
9 Apr
Regular lectures, not video!  Quiz after Friday's lecture. Language bridges between social cognition, learning, and is core to human ethics.
9
Emotions & Intelligence; Mind & Consciousness;
Lecture 13; Lecture 14
Code for Coursework 3 due Monday.
Labs mandatory!
Tuesday quiz 13, 14

16 Apr
Lab: Round-robin game competition; if you can't make your lab you need to get someone else to make your code run for you (NOT the tutors!)
(Video) lectures (with normal quiz Tuesday) take a functionalist, empirical, science-driven perspective on topics critical to human society and ethics, in preparation for considering AI's role in ethics.
10
Ethics & Society; Politics & Regulation
Lecture 19; Lecture 20
Coursework 3 writeup due Monday
Beauty Competition Tuesday's Lab; Final games competition Wednesday's Lab.
Robots must be disassembled by this week.
Unless they win the beauty competition!
23 Apr
Labs this week are the (optional) beauty contest & game championship and the non optional disassembly of the non-winning robots (mandatory for undergrads; MScs can hold on to their robot if they are using it for CW4.)
Regular lectures, not video! (But probably both early in the week.)  This material probably affects your coursework least, but your career (and life) most, of any of the lectures. 
Quiz Friday, 19, 20
11
Revision or Guest Lecture
30 Apr
No more quizzes!
Rev
 Revision or Guest Lecture

7 May
Coursework 4 due for postgraduates.

If you want to meet with Dr. Bryson privately, schedule an appointment during her office hours; if there are none coming up, email her.  Please don't email lecturers or tutors questions about lectures, course content, or coursework.  Questions are good work done by good students, both they and answers belong to the whole course and should appear in the Moodle forums.  It's also easiest to get our attention in lecture, lab and the moodle forums.  Note that extension requests go to your Director of Studies, not to Lecturers.  The best time to talk to Dr. Bryson is often immediately after lecture (before is OK iff everything technical seems to be set up OK.)

Is there a text?

This is an advanced course for which there is (so far) no single text book.  One of the skills taught in this course is reading primary literature–that is, you should be able to read some of the publications cognitive scientists use to communicate to each other.  However, we have also ordered a large number of books on robots, AI, LEGO robots & Cognitive Science for the library.  Have a look on the shelves or in the catalogue.  Let me know if you particularly like any of these books and/or think we should get more copies.



page author: Joanna Bryson