Month: August 2010

Top news cites referenced from Wikipedia

Posted on

Jodi Scheider has just alerted the wiki research mailing list about Ed Summers’ recent statistics on outgoing links from Wikipedia: The top hosts referenced in Wikipedia.

NIH with PubMed is the big one, followed by the DOI resolver site, IMDb and Google Books. Surpricingly you will find as number five! I wonder if it has anything to do with his Find link Wikipedia web service.

Almost a month ago I changed my old Perl program that I used for the Scientific citations in Wikipedia paper to look into the ‘cite news’ template instead and examine the outgoing links from this template. Such an analysis will look specifically on news citations. On the top twenty list for an 2008 English Wikipedia dump I found:

 1 14443
 2 3224
 3 2729
 4 2675
 5 1838
 6 1781
 7 1584
 8 1443
 9 1420
10 1278
11 1198
12 1119
13 1026
14 1006
15 1005
16 987
17 975
18 969
19 913
20 904

The topper is also the top news site in the list that Ed Summers created. I expected BBC, New York Times and Washington Post to be high on the list, but not that BBC would be so well ahead of the others. The Australians are coming along well, and their first is Sydney Morning Herald

I get The Hindu as 39, Haaretz 62, while English Al Jazeera is down on number 100. The most template-linked Danish news source is Politiken as number 245 with just 82 citations, ahead of French Le Monde, but behind the Swedes’ Dagens Nyheter.

Danish Information incorporates data from Wikipedia by Johannes Wehner afaik but gains only 2 citations. Politiken has some original English content, while Information has none afaik.

Dusting off an old artificial neural network code

Posted on Updated on

During my Master Thesis I constructed a ‘standard’ two-layer artificial neural network. It was later expanded and added to our Lyngby Toolbox, that is a Matlab toolbox for analysis of functional neuroimages (and other things you can think of).

Today I heard that Morten Arngren was working with a neural network with pruning and regularization optimization, and I mentioned our Lyngby code.

Now I just wanted to see if it still worked alright, but saw that it lack a small example to get people started. So I made a small noiseless one-dimensional example with a sine wave:

X = 3*rand(50,1); T = sin(X);
[V,W,EAcc,Info] = lyngby_nn_qmain(X, T, 'genoptim', ...
'EarlyStop', 'trainset', 1:25);
X0 = (0:0.1:3)';
[Y, H] = lyngby_nn_qforward(X0, V, W);
plot(X,T,'o', X0,Y,'-')

Here the generalization is just controlled with so-called ‘early stopping’ using the first half of the data set as a training set (25 data points) and the second half for ‘validation’. The forward function generates the output (Y) of the neural network.

The more complicated and longer running ‘Pruning2DRegGridSearch’ option both prunes and optimizes two regularization parameters (ridge regularization parameters): one for the input layer of parameters (weights) and one for the output:

X = 3*rand(50,1); T = sin(X);
[V,W,EAcc,Info] = lyngby_nn_qmain(X, T, ...
'genoptim', 'Pruning2DRegGridSearch', ...
'reg', [10^-5 10 10^-5 10], ...
'trainset', 1:25, 'info', 1);
X0 = (0:0.1:3)';
[Y, H] = lyngby_nn_qforward(X0, V, W);
figure, plot(X,T,'o', X0,Y,'-')
figure, lyngby_nn_plotnet(V,W)

This latter example plots the output and also the optimized neural network structure (see the figures).

In this case the one-dimensional example presents no problem for either approaches. When I used it for funtional magnetic resonance imaging (fMRI) data in my thesis and in the article Plurality and Resemblance in fMRI Data Analysis my recollection is that neural network optimization was a long-winded process.

The Lyngby Toolbox can be downloaded from here.


Remember the big Ubuntu 10.04 Freeze back in 2010?

Posted on

My Ubuntu computer has begun to freeze out of nothing. It has been happing since I installed Ubuntu 10.04 LTS. I am not the only one: A Ubuntu 10.04 (Lucid Lynx) Random Freeze / Hang-up Ubuntu Forum thread has since 10th May 2010 gained 881 entries (August 2010 count) and the Ubuntu 10.04 random freezes Launchbad thread reaches 165.

Keyboard, mouse, screen and Ctrl+Alt+Backspace freezes and only a hardware reset seems to be available for getting me back. The frequency is around once a day or more often. It is not possible to reproduce. There is nothing to look for /var/log/kern.log, /var/log/messages or /var/log/syslog. The upgrade to 10.04.1 did not fix it. It has been happening with Linux 2.6.32-24-generic kernel and previous versions. It is happening on an older Dell laptop which I recently installed with 10.04. I have another Acer Aspire One with Ubuntu 10.04 that does not have the problem. It was originally installed with 9.10. I doesn’t seem to be tied to a specific use of an application. “System > Admin > System Testing” does not led to any useful crash or information.

There have been loads of suggestions on the Internet for why this is happing:

  1. Bad RAM. But it worked before?
  2. The new ext4 file system. And, yes my file sytem ext4 (see with df -T). Before it was ext3. So what do I do?
  3. Nvidia driver. Well the graphics card is Intel 945GM/GMS/GME, 943/940GML.
  4. 64 bit issue. No, I am running 32 bit.
  5. Compiz with all its flashy effects. Perhaps.
  6. ACPI. Perhaps. databubble suggests “noapic nolapic acpi=off” as kernel parameters during boot. When I introduce “acpi=off” my mouse doesn’t show up in X… Wow what a bug.
  7. Etc.

In the process you learn about Alt+SysRq+REISUB, i.e., “busier” backwards. This will actually reboot the freezed computer. As bugfix users suggest going to Debian and PCLinuxOS. One user writes: “After 4 years of Ubuntu and 3 months of random freezing I’ve migrated to PCLinuxOS.”

I am presently limping in Ubuntu 8.04 and looking on Debian. This is a show-stopper for Ubuntu.

(Typos corrected 4th September 2010)