Eldan Goldenberg's lab notebook

Entries for February 2005

Graphics in C++

I'm trying to put together a simple graphical visualiser for the simulator I'm writing. All it needs to display are a rectangular arena, a couple of circles for the agent and the object it's tracking, and some lines to represent the agent's field of vision. It seems to me that this ought to be pretty easy, but I know nothing about graphics programming at all, so I'm looking for a library that I can use to simplify the process.

The problem is that I'm working on a Mac, and so far the only things I've found are very low-level, and entail a lot more work than makes sense when my requirements are so basic. I did find one library that looked ideal and had the added bonus of claiming to be cross-platform, but it turns out to require X Windows, and I couldn't get it to work with the X terminal that's built in to Mac OS.

Can anyone point me to a suitably high-level library for Mac OS or (even better) that will give me cross-platform graphics?

Deadlines, deadlines

The deadlines I'm currently working to; some self-imposed and some external:

This week

Have my simulator basically up and running so I can start testing it

Before Spring Break

Have my simulator well enough tested that I feel I can trust its output to reflect what's really going on in it. Have a good enough outline & division of labour for the Notacon talk that I'm ready to start writing content.

During Spring Break

Gather data from the simulator (I'll be skiing, so I want to be able to set up an experiment each evening and let it run for 24 hours). Write a first draft of my parts for the Notacon talk.

April 4^th

ECAL submission deadline. I have a separate project I need finished for this (not the simulator I'm working on right now, but one similar enough that I'll be able to re-use a lot of code). I really don't know if I can make this or not, but I'm going to try.

April 5^th & 7^th

Present a paper in the Computational Neuroscience seminar. It's 25% of the assessed grade for the course, so it had better be good.

April 8^th - 10^th

Notacon

Mid-June

Defend thesis proposal, so I can move to Seattle knowing I'm on the right track with work. [this is what the simulator I'm currently working on will contribute to, which is why it's still my top priority even though the deadline's longer than other things]

It's going to be a busy few weeks coming up.

When text hampers communication

Last year I gave my first conference presentation. It was also a first in another way, because it was the first time I had given a presentation that was really well-rehearsed, had gone through successive drafts, and had a structure that had been picked over carefully.

Continue reading "When text hampers communication"

Notacon submission

Sean has agreed to co-present with me at Notacon, which will be good. I think between us we'll do a better job of holding an audience's attention for 30-60 minutes than I would have been able to do alone. We submitted our proposal yesterday evening; here is the talk outline:

"Walking before we run: how biology is inspiring progress in AI"

Computers may be outstanding tools for processing data, but they still require constant human oversight and perform poorly at many tasks that come naturally to even ‘lower’ animals.   Meanwhile our attempts at making computers do these sorts of tasks are giving biologists and psychologists new insight into how natural systems work.

This talk will show you examples of recent work in artificial intelligence that draws its inspiration from biology, and life sciences research that is informed by computer models. We hope to convince you that these are useful for everything from spam filtering to Mars rovers to diagnosing heart failure. 

Suggestions for content are welcome. At this stage I have only a pretty vague idea of which examples to give and how to tie it together as a whole.

update: the proposal's been officially accepted.

Trick the birdie

I'm putting together a talk proposal for Notacon. In the process, I'm finding myself jumping ahead (all they want this month is a 150 word abstract) and thinking about things that should actually be in the talk. One thing that should definitely be included is an example of how biological evolution can sometimes produce highly specialised controllers that are easy to trick, just like artificial evolution. It should only be a small anecdote, but it's starting to bug me because I know the specific example I want to give, but I can remember too few details of it. It's something that was described in a lecture back when I was an undergrad, but seeing as that was last millenium my memory of it is a little hazy on details. So here is the phenomenon:

Continue reading "Trick the birdie"

Why a work blog?

I'm starting a blog that will just be for writing about my work. There are really three purposes this should serve:

• Practice writing about what I do, for a technical audience broader than just the people in my lab or at the more specialised conferences in my field. Ultimately my work will be a lot more worthwhile if I can go to a psychology conference and convince that audience that they can use the data I've gathered from computer models, but this calls for a different sort of communication skill than going to an AI conference where everyone is familiar with the basic assumptions we make.
• Hold myself to account by giving myself set targets with work that's free-form enough that it's far too easy to drift for a few weeks at a time, and reporting on progress. This is especially important to me because I'll be collaborating remotely with my advisor for a few years, during which time I won't be having weekly lab meetings and such things to keep me on track.
• Hopefully get some discussion going to get feedback on half-baked ideas and pre-draft versions of papers.

At some point soon I'll post a brief about me / about my work introduction, but for now I just want to get started. For now here is a rather out of date page for my research group, which will soon be replaced, but at least gives some background: the Dynamics of Adaptive Behavior Lab.