I have been thinking about going to grad school in computer science. I thought it will be nice if the program is open about using Lisp as a research tool (preferably in machine learning). Do you guys know of any?
HalfEmpty wrote: > I have been thinking about going to grad school in computer science. I > thought it will be nice if the program is open about using Lisp as a > research tool (preferably in machine learning). Do you guys know of > any?
HalfEmpty wrote: >> What made you lose interest in computer programming?
> I think there is a pretty good overlap between computer programming > and computer science.
Don't say you weren't warned.
> If you know of a computer programming graduate degree that uses Lisp, > I'd appreciate that too.
Sorry, can't help. But as long as you are not listening: just ask them if you can use Lisp. If they do not say, "We do not give a rat's ass what you use."*, find another program.**
hth,kt
* It does not have to be a rat.
** On second thought, just asking this question will cost you admission to the best programs.
På Fri, 09 Jan 2009 05:34:23 +0100, skrev HalfEmpty <ToPacifythe...@gmail.com>:
> I have been thinking about going to grad school in computer science. I > thought it will be nice if the program is open about using Lisp as a > research tool (preferably in machine learning). Do you guys know of > any?
I can't quite see why. Machine learning these days (at least the bit's I have researched) seem to focus on Bayesian logic or neural networks and to some extent genetic algorithms. In these you are usually better off with a linear algebra oriented language like mathematica or matlab. There is a interesting approach to gaming that uses feedback from sucessfull moves that is suited for lisp. (reinforcemet learning)
HalfEmpty <ToPacifythe...@gmail.com> writes: > I have been thinking about going to grad school in computer science. I > thought it will be nice if the program is open about using Lisp as a > research tool (preferably in machine learning). Do you guys know of > any?
Don't know about machine learning, but if your interests include computational/system biology, check out those projects:
On Jan 8, 8:34 pm, HalfEmpty <ToPacifythe...@gmail.com> wrote:
> I have been thinking about going to grad school in computer science. I > thought it will be nice if the program is open about using Lisp as a > research tool (preferably in machine learning). Do you guys know of > any?
I went to grad school in machine learning and used mostly Lisp (e.g., http://www.cs.berkeley.edu/~bhaskara/alisp/). A significant fraction of Phd advisors at top schools will be hands-off enough not to care what language you use. Just apply all over and ask before you go. I wouldn't make this your top criterion though.
Btw, in case people bring up the lack of numerical/linear algebra libraries in Lisp, note the existence of Matlisp. - b
On Thu, 8 Jan 2009 20:34:23 -0800 (PST), HalfEmpty
<ToPacifythe...@gmail.com> wrote: >I have been thinking about going to grad school in computer science. I >thought it will be nice if the program is open about using Lisp as a >research tool (preferably in machine learning). Do you guys know of >any?
The Natural Language Processing group at the University of Oslo (Norway) uses Allegro CL heavily, and has been focusing increasingly on machine learning techniques in recent years. The machine learning technology is usually implemented with third party tools not implemented in Lisp though. You can check out some of the projects they participate in at http://www.delph-in.net.
They offer both Masters degrees and the occasional PhD stipend.
My own Masters thesis investigated machine learning models and included a fair amount of CL coding (along with lots of C++). I think it would be quite possible to do a CL heavy thesis there if you're so inclined.
On Jan 8, 9:34 pm, HalfEmpty <ToPacifythe...@gmail.com> wrote:
> I have been thinking about going to grad school in computer science. I > thought it will be nice if the program is open about using Lisp as a > research tool (preferably in machine learning). Do you guys know of > any?
> I went to grad school in machine learning and used mostly Lisp (e.g.,http://www.cs.berkeley.edu/~bhaskara/alisp/). A significant fraction > of Phd advisors at top schools will be hands-off enough not to care > what language you use. Just apply all over and ask before you go. I > wouldn't make this your top criterion though.
That's good to know.
I'm actually not worried too much about using Lisp. However, I'm worried that I didn't have a bachelor degree in CS (it was Physics), and I have never taken any course in CS. I taught myself everything by chewing through textbooks, dissertations and journal papers. I don't know how seriously top grad school would consider me, since none of that can be put on paper.
I did a good amount of work in Genetic Algorithm and Neural Network, which were heavily influenced by the work done by Kenneth Stanley (Neat and HyperNeat).
Right now my interested lie in hidden variable Bayesian Network and reinforcement learning (Alisp looks cool. I will definitely check it out.)
On Sat, 10 Jan 2009 13:15:56 -0800, HalfEmpty wrote: > I'm actually not worried too much about using Lisp. However, I'm worried > that I didn't have a bachelor degree in CS (it was Physics), and I have > never taken any course in CS. I taught myself everything by chewing > through textbooks, dissertations and journal papers. I don't know how > seriously top grad school would consider me, since none of that can be > put on paper.
If I were you, would not worry about that. Especially if the people who are writing your recommendation letters can talk about it briefly, or you can show previous work directly (in the form of results, papers, etc).
I am in an econ PhD program. Our department is known to accept people who didn't major in economics (but did math, physics, or even if they majored in something totally remote but have taken a good amount of math courses). The attitude is that you will be taught most of what you need to know in their courses anyway. Good graduate programs are interested in attracting smart and motivated people, and what you did in college is of secondary importance.
Also, in answer to your original question: it seems that you are worried about something that should not come up in practice. They will teach you theory and methods, and should allow you to pick the language in which you implement them. When you write papers, people care about your results, and the language you used for programming them is not even mentioned most of the time (unless the language was important in getting the results, etc).
On Jan 10, 1:32 pm, Tamas K Papp <tkp...@gmail.com> wrote:
> On Sat, 10 Jan 2009 13:15:56 -0800, HalfEmpty wrote: > > I'm actually not worried too much about using Lisp. However, I'm worried > > that I didn't have a bachelor degree in CS (it was Physics), and I have > > never taken any course in CS. I taught myself everything by chewing > > through textbooks, dissertations and journal papers. I don't know how > > seriously top grad school would consider me, since none of that can be > > put on paper.
> If I were you, would not worry about that. Especially if the people > who are writing your recommendation letters can talk about it briefly, > or you can show previous work directly (in the form of results, > papers, etc).
Yeah, assuming your GRE scores and GPA are above threshold, rec letters are what people look at. If you can find some way to get one or two CS professors to write you strong letters (e.g., by doing a research project with them), that will help a lot. If your previous work resulted in a conference paper that would be another plus. If not, you have eleven months before the next round of grad school deadlines, which is plenty of time :) - b
> On Jan 10, 1:32 pm, Tamas K Papp <tkp...@gmail.com> wrote:
> > On Sat, 10 Jan 2009 13:15:56 -0800, HalfEmpty wrote: > > > I'm actually not worried too much about using Lisp. However, I'm worried > > > that I didn't have a bachelor degree in CS (it was Physics), and I have > > > never taken any course in CS. I taught myself everything by chewing > > > through textbooks, dissertations and journal papers. I don't know how > > > seriously top grad school would consider me, since none of that can be > > > put on paper.
> > If I were you, would not worry about that. Especially if the people > > who are writing your recommendation letters can talk about it briefly, > > or you can show previous work directly (in the form of results, > > papers, etc).
> Yeah, assuming your GRE scores and GPA are above threshold, rec > letters are what people look at. If you can find some way to get one > or two CS professors to write you strong letters (e.g., by doing a > research project with them), that will help a lot. If your previous > work resulted in a conference paper that would be another plus. If > not, you have eleven months before the next round of grad school > deadlines, which is plenty of time :) > - b
Top universities regularly admit students into CS grad schools even when their formal training is from another discipline. What counts is the ability to demonstrate your preparedness for graduate work in CS, whether through formal study or other means. The fact of your motivation for self-instruction is a strong plus, which some regard more highly than certain modes of formal training.
Demonstration of your preparedness for graduate study in CS may include the GRE in CS or math (which I believe includes topics such as AI and theory), but more importantly includes your interests in pursuing research related to the field. In this regard, the suggestion to submit evaluations from a CS professor (or two) acquainted with (whether as supervisors or as collaborators, or colleagues) some of your research is an excellent one. Your evaluators need not be professors, they can be professionals established in the field. Back up these evaluations with a published research paper. This is what you need to demonstrate preparedness for graduate work in CS. If you want to work in Lisp, make sure you can show code that applies.
In any case, once a graduate school is suitably impressed with a candidate, it is normal procedure to require the completion of certain core CS courses in order to make up for deficiencies in background. Your BS in Physics is not an insurmountable concern.
