This document is not linked anywhere yet
I am still considering this plan in relation with my original dissertation proposal.
(for more information on all this, see my current dissertation proposal)
When it comes to study, I am a utilitarian realist and I'm painfully aware of my cognitive limitations. This is why I am studying Cognitive Science. In order to improve my own [memory, decision-making, attitude, etc.], I have to know how human learning works. The best formal approach (that I could find), to that sort of enquiry is Cognitive Modeling. My goal in researching, though, is not knowledge in itself, but the application of that knowledge in physical or abstract cognitively efficient tools. See my research interests, as outlined in my application to the Carleton University Ph.D program.
Unfortunately my interests are too broad and spread over a large set of domains (life's interesting because it's diverse). But I have to keep a realistic target:
- For my dissertation I have to pick one topic (tough choice), and follow it in a combination of depth (PhD) and breadth (Cognitive Science) to my best abilities. As you can see, a PhD in Cognitive Science does have an interesting contradiction built in.
- The fact(s) that I'm not Canadian (yet?) and that I have the additional responsibilities for my family, strongly reduce my chances of establishing an academic presence (writing formal papers for publication, attending conferences, indulging in extensive empirical studies, even socializing with my colleagues)
I am in this CogSci program for its practical applications, not for an academic career. So I shall make my dissertation topic choice based on that. I will concentrate on MonDoc (the set of principles which outline my own approach to communication), and the applications thereof.
That does not mean that I don't see the place for theory. Not at all. Theoretical work is supposed to lead to practical realization. By themselves any of the two are doomed do fail one way or another.
So here's my PhD plan (contingent on approval by my supervisor):
- 32.534 or 32.535 Tutorial in Selected Problems of Philosophy. Read up on the philosophy of representations (originals, as much as possible, because I keep getting confused by second- or third-hand interpretations of their writings): Locke, Wundt, Hume, Russell, Quine (web of scientific belief), Pinker (mind works), Fodor (no it doesn't), Wittgenstein-Davidsoon-Dummett-McDowell (all thought requires language) ,Dretske (), Millikan (), Kosslyn-Pylyshyn (mental imagery), Putnam (non-algorythmic processing), Clark (DST) and whatever else my more philosophically inclined friends will suggest
- in the meanwhile (as I've been doing during my "Comp"s), continue gathering evidence for my MneMonic model of cognitive architecture, identify potentially solvable disputes and prepare empirical tests (experiments) to attempt to solve them
- polish up my set of MonDoc principles and add the necessary argumentation for each principle
- work on the MonDoc engine itself and generate as many applications as I have time for
- if all of the above goes smoothly, integrate self-monitoring, automatic consistency preservation, suggestions and maybe even NL input: add MneMonic to MonDoc, resulting in an intelligent interface (literally, not the metaphors that are used to describe some of the existing self-adjusting interfaces)
(for more information on all this, see my current dissertation proposal
Other courses I'm still looking forward to:
- 95.506 and 95.555 Natural Language Processing,
(I think Edina mentioned something about one of these)
- 95.510 Topics in Artificial Intelligence,
- 95.520 Cerebral Computation,
- 95.581 Topics in Machine Learning.
- 95.510 Cognition and Artificial Cognitive Systems,
- 49.530 Perceptual Processes,
- 49.570 Experimental Research in Cognition,
- 49.624 or 49.625 Neuroscience Techniques,
(this may be what I'm doing for my Neuro Comp?)
- 29.561 Cognition and Language
(which I tried before as 07.503)
Here's a couple of words about my Comps (all directed toward what I'm interested in).
- Psychology (with Gitte Lindgaard) - MonTaGe (an application of MonDoc for Cognitive Work Analysis)
- Neuroscience (with Bruce Hutcheon) - Microscopy in rat brain slices (and hunting down learning patterns in the literature)
- Computer Science (with J.P. Corriveau: TBD) - MonDoc engine and MonTaGe (using the research on CWA, create a system design and do the actual OO implementation of MonDoc, with MonTaGe on top of it.)
- I'll also take a Directed Studies with Andy Brook (TBD) on "Representations from several points of view" that would bring a more theoretical look at Knowledge Engineering than many Computer Scientists don't seem to take into account.
I hope this whole page makes it clear that I'll work interdisciplinary :) I'm hoping to get
- Robert West as my dissertation supervisor.
I would use:
- Cognitive Modeling (MneMonic - for the user model),
- Philosophy (Representations - for the theoretical background),
- Psychology/HCI (MonDoc - for the user interface),
- Computer Science (MneMonic and MonDoc - for the system(s) design, actual implementation)
... and maybe
- Cognitive Modeling/DST (because I want to explore mathematical models of learning),
- Linguistics (if I use the text analysis application of MneMonic as a spelling/grammar checker) and
- Neuroscience (if I get to find some more neurologically plausible learning parameters to apply in the neuronal layer of MneMonic).
So in the commitee we may have any combination of:
- Andy Brook (Philosophy and Linguistics)
- J.P. Corriveau (Computer Science and Cognitive Modeling)
- Paul Hirchbuhler (Linguistics)
- Bruce Hutcheon (Neuroscience)
- Bruno E (Cognitive Modeling and HCI)
... other local or UO people you may suggest?
For the external examiner(s), we may pick
- Boicho Kokinov (NBU - Cognitive Modeling and Psychology - since he knows about both MneMonic and MonDoc)
- Maxim Stamenov (NBU - Linguistics and Philosophy - since he knows some of my work)
... and there are many biggies whose work I admire and whom I'd like to drill me on this (like Hofstadter, Gentner, Holland, Clark, Kosslyn, Knuth, etc. - yeah, Radu, dream on :) You may not see why all these people (and many more) had an influence on the development of MonDoc and (especially) MneMonic, but I assure you, they did.