Adrian Brasoveanu

Fall 2014: Plan

Tue 02 September 2014 by Nate Arnett, Deniz Rudin, Clara Sherley-Appel, Adrian Brasoveanu

The four of us (and some other email-silent partners :-)) hatched a plan to have an ACT-R centered series of meetings this fall. The plan is to start by reading Anderson’s most recent book:

  • Anderson, John R. 2009. How Can the Human Mind Occur in the Physical Universe?

and then work our way back to:

  • Anderson, John R. and Lebiere, Christian J. 1998. The Atomic Components of Thought

and even all the way back to:

  • Anderson, John R. 1990. The Adaptive Character of Thought

These books are fairly cheap on amazon (get used copies of the older ones…).

We will interweave this with learning how to implement the main models in Python ACT-R:

Check out the tutorials in particular:

We could then take a look at ACT-R in linguistics (the work of Lewis & Vasishth, Hale, and various others) and think about starting some projects together given our theoretical knowledge of the framework and also the computational know-how we will acquire with Python ACT-R.

For example, Nate will talk to us about the Lisp implementation used in the Vasishth 2008 paper on polarity processing (a link to this implementation seems to be available online), and possibly about some ACT-R related R code (but we need to look into this a little bit first and ask for permission from the original authors).

For the people that need to develop their Python skills, a book you could start with is this one:

but there are many other free online resources: books, tutorials etc. They’re all just a google search away. You could also consider enrolling in this coursera course that starts on Sept. 15:

Adrian knows other courses taught by the same instructors (but not this particular one) and they are very good. There are other courses on Coursera and on other MOOC platforms out there on Python, R, stats, scientific computing etc. that you could also take a look at. This type of online education could be a great way to acquire the basic technical (math/comp sci) background knowledge you need to do this kind of more computer/math intensive research. We can cover some of the prerequisites in our meetings, but we’ll probably be much better off if the meetings are dedicated to the specific linguistics and cog sci issues we are primarily interested in rather than to general prerequisites of this sort.

Clara also mentioned that there is an online facebook group going with some alumni/ae in the industry where you can post code for your projects and get critiques/help. It hasn’t been super-active on the learners’ end, but our industry folks have been very helpful whenever someone does post their code. If you’re not on invite for that and would like to be, just let Clara know.

When & where: Tuesdays, 11 am — 12 pm, The Cave. The seminars scheduled Tuesdays end at 10:45 am or start at 2 pm. On occasion, we might start a bit after 11 am depending on the exact schedule of the preceding seminar (which obviously takes precedence).

The first meeting: Tue, Oct. 7, 11 am — 12 pm. Let’s plan on reading the Anderson 2009 (How Can the Human Mind Occur in the Physical Universe?) book for Oct. 7. It’s a short book. We’ll discuss as much of it as possible during the meeting and make a specific plan for the next meetings.

Enjoy the last month of this extra long summer break and see you then! We’ll probably send a reminder on the LaLoCo mailing list closer to the date (if you want to be added to that mailing list, please get in touch with Adrian).