Contents || Introduction || Background || Research || Group || Publications|| Demonstrations || Teaching || Interests || Links || Contact || email
Probably Informative are the web pages of the Edinburgh Machine Learning Group.
"Hacking" is an illegal activity. This excellent article entitled "Is your Child a Hacker" gives you top advice on how to spot if your child has taken up hacking. Especially look out for copies of "Programming with Perl" ...
I am editor of the Open Directory Project (DMOZ - used by Google for the Google Directory) Computers: Artificial Intelligence: Neural Networks: People category, which I aim to make as comprehensive as possible. Please submit a homepage if you are a researcher or worker in this field and are not included (no guarantees can be made for inclusion: Dmoz has strict guidelines for what can and cannot be added).
I am editor of the Open Directory Project Computers: Artificial Intelligence: Conferences and Events category and a number of other related conference categories. Conference Details can be found there. I am an organiser of a National E-Science Centre workshop on Scientific Data Mining, Integration and Visualization (SDMIV).
The key to good images and good compression is knowing when to use what approach for what. My rule one: avoid using gifs unless you have to (yes I know I've already broken that one). I find that anything that gifs were good for, PNG does better, and using png doesn't involve propagating a stifling unisys/compuserve patent. In general, despite what most webpages suggest, jpeg is not good for everything! It seems that jpeg is reasonable for natural truecolour images, whereas png is much better for diagramatic, cartoon or presentation related images. If on the other hand you have images of varying type or mixed type and pictures (such as scanned documents), then djVu is for you.
This is a fun business unless you are engaged in it! Much of the year central Edinburgh properties move, and move fast. We had bought our flat about 20 hours after it went on the market - and 8 of those we were asleep. Looking for a purchase? Best forget getting any work done and simply go to Edinburgh Solicitors Property Centre and continuously hit refresh - you want to be the first to see a property come onto the market don't you?
So I don't like emacs. Yes I know I am supposed to like emacs, but I don't. Why? Because I am used to editors such as WinEdt, and frankly, in comparison, I find it is always hard to get emacs to do what I want it to (although perhaps some people like spending years writing lisp - might as well write your documents directly in lisp IMO). Yes WinEdt costs money (but not much). No WinEdt is not GPL. But it is good, excellently supported, well priced, and does what I want. The main problem with it is that it is only available for windows. Linux version on the request list, Alex! :-)
Many people will say that LaTex is a must if you do any technical writing. Even if you are stuck with windows MikTex does the job. And it integrates into WinEdt. For those still using MS software, IMO it is well worth investing the time into learning Latex.
One annoying feature of latex is the verbosity of getting bold characters in math mode. Generally latex users develop their own set of shortcuts for doing this. I have defined all the shortcuts that I use (for the whole roman, greek and amssymb letters that are available in latex), and put them in to shortbold.sty. Then commands such as \xB produce a bold x, \YB a bold Y. Likewise \alphaB, \PiB, \digammaB do what you would expect them to.
Matlab is a handy tool for initial analysis, although it has serious deficits for large problems. Another problem it has is the lack of a single function for loading variable length data, or data of mixed type. I wrote the mfile loadcell.m to fill that gap for small files.
br>