Searching the Matrix

Here is an interesting concept on the Google distance, basically using Google to figure out distances between words. It was devised by Cilibrasi of CWI who I mentionned earlier in this entry. In reference to this, I eventually mentionned that you needed to use this concept to search through videos. Incindently, the new Google search engine now goes through videos but uses the text of the show for the blind to search through them. This is a nice trick, but we will need tools like that of Cilibrasi based directly on the video distance to figure out exactly what type of movies/scene we are looking for. It all looks like the convergence of several media are going to make this technology happen.

Colloque "Mathematiques du Reel" / Colloquium on Real World Mathematics

Les inscriptions pour le Colloque "Mathematiques du Reel" sont ouvertes. Le Colloque est gratuit mais il faut s'inscrire. Cela se passe du lundi 21 au vendredi 25 mars 2005 dans les locaux de EdF R&D, 1 avenue du Général de Gaulle, 92141 Clamart, Salle Pierre Ailleret. Pour plus d'information sur le programme et les presentateurs peuvent se trouver surle site du colloque.

Registration for the Colloquium on Real World Mathematics is open. The colloquium is free but one has to regsiter. It will take place Monday through Friday March 21-25. The location is EdF R&D, 1 avenue du Général de Gaulle, 92141 Clamart in the Pierre Ailleret room. More information on the program and the speakers can be found here (the meeting will be held in French.)

What's the difference? You're their all-time best seller

I was listening to NPR the other day and heard about this book on words that are extremely difficult to translate from one language to the other. One of the example the author found for French was "esprit de l'escalier" which I did not recognize right away, but eventually remembered to be indeed "l'esprit de l'escalier", you know, the comeback thought that crosses your mind and would have been extraodinarly a propos 2 minutes before in the heated discussion you just had but did not have the wittiness to find, that's having " l'esprit de l'escalier" or the wit as you march down the stairs.

It's all downhill from here

In a recent communication to a conference of the American Heart Association it was reported that hiking downhill removed blood sugars and improved glucose tolerance, while uphill hiking mostly improved levels of fats called triglycerides. So the real question is: do you remove blood sugar because your brain works more when you go downhill than when you go uphill or is there another mechanism at play...

[Update September 2008: an explanation is given in the comment section: "...When walking down the mobilization of glucose from liver or muscle glycogen is low and if there is still excess supply of glucose, it is stored in the form of glycogen or fat (triglycerides by example). When walking uphill mobilization of glucose is high, it can be done from the reserves of glycogen (fairly weak in the muscle and a little more important in the liver) at first and then triglycerides from fat tissue which in order to reach the muscle cells must pass through blood (hence increasing the blood level of triglycerides).." ]

Excellent

Here is a traditional inspirational poster on Goals and the Despair one on the same subject.

What's in a name ?

This one is called a waterspout


But when it hits downtown Miami, it's a tornado

AI and Human Cognitive Problems

I have been invited to make a presentation at the tenth anniversary of the SCM company. The title of my talk will be the following :

"L'Intelligence Artificielle peut-elle guider un véhicule autonome dans le désert ? peut-elle aider à diagnostiquer des problèmes cognitifs chez l'homme ?"

or in English

"Can Artificial Intelligence guide an autonomous vehicle in the desert ? can it help in the diagnosis of cognitive problems in humans ?"

It will reflect some of the current issues I have identified so far in building an algorithm for our entry in the DARPA race and inversely how I think some of the techniques I have seen developed in the Artificial Intelligence community could be of help in the diagnostics of cognitive problems in kids and grown-ups. The latter subject is fascinating because there is seldom a communication between the engineering world of robot building and the world of applied psychology.

I could have talked about other subjects where I am far more competent (space, nuclear,...) but I am not sure that finding solutions in these areas would be a substantial contribution. I am not quite sure there are Grand Challenges in those areas anymore.

Darmstadt, Huygens has landed

It looks like Huygens has landed/survived and already beamed about 1 hour and half of data to Cassini. Cassini will beam back data to Earth at about 10 AM CST today.

Star Navigation and Image Registration of Proteome Gels

It would seem to me that the automatic registration of images of proteome gels could easily use the techniques we use for satellite navigation with star trackers. The author of this paper, Douglas Kell, also seems to be a discoverer of a Morse code type of communication between cells.

Twister Twister Land

On Dec 28th we had this warning

628 WFUS56 KLOX 290643 TORLOX CAC037-290730- BULLETIN - EAS ACTIVATION REQUESTED TORNADO WARNING NATIONAL WEATHER SERVICE OXNARD CA 1034 PM PST TUE DEC 28 2004 THE NATIONAL WEATHER SERVICE IN OXNARD HAS ISSUED A * TORNADO WARNING FOR... SOUTHWESTERN LOS ANGELES COUNTY IN SOUTHWEST CALIFORNIA INCLUDING PALOS VERDES...SAN PERDO...ROLLING HILLS... TORRANCE AND CARSON * UNTIL 1130 PM PST

Then this happens today..

or this


It looks like this movie much too much for my taste

My tricks to avoid cognitive overload

We are all swamped with E-mails and so forth. Here are some of the tools I am using to cope with some of these issues:

- First, I need to be able to store and search large amount of data. I have a gmail account. Let me know if you want one. The nice thing about it is the 1GB quota allowed. It is web based, no need for a local computer to have access to your E-mail and you can search through all your E-mail stack using the google search engine technology. There is also a Gmail notifier that tells you when a new mail has arrived. Nice features.

- Second, I want to know what's new from websites I care about. I use Watchthatpage, a free service that monitors the web pages I am interested in. I ask them to send me an email on my gmail account once a day and tell me what is new in these sites of interest. No random surfing anymore. You can have several e-mails sent for different subjects of interest. Right now I have about forty sites on my watch list.

- Third, I need to search through my HD in a manner that is faster than the stupid search feature in Windows. For this I use the Copernic desktop seach. I used the Google Desktop search engine, but it did not index pdf files. I have too many of these pdf files that I have to resort to Copernic and it works pretty well. I'll switched back to the Google desktop search as soon as it indexes pdf files.

- Fourth, I need to defend myself from worms and other viruses. I use the Kerio firewall. I also use Ad-Aware, Spybot Search and Destroy and the new Microsoft Beta anti-spyware softwares to clean up my computer. I also use a web browser that has less issues with security than Internet explorer. I either use Firefox or Opera. I also connect automatically to Windows Update.

- Fifth, I need to download some of my thoughts in a blog. I use blogger, you set up an account and can have different blogs under the same account. If I know you, let me know if you start one.

- Sixth, I need to "see" what part of my harddrive is full. The only way to do this efficiently is to use Spacemonger.

- Seventh, in order to see all the pictures on my HD, I use Picasa.

- Eighth, I use sitemeter a free counter that tells me who comes to my web pages and see the interest they may raise.

- Ninth, in order to organize my thoughts, I use Freemind that can export the graphics into an html page. Very nice. I am also beginning to use GraphViz. I use the Python programming language among others.

- Tenth, sometimes I need copy entire section of a website in order to browse it offline. I use HTTrack Web Copier.

All these softwares are free. Enjoy.

We sleep because we want to assimilate bimodal priors

Pierre sent me a pointer to a Time article on why we sleep where they seem to talk about the latest research showing that it is not because of your body but rather your brain that you sleep.

In the PBS/Nova story on the secret life of the Brain, they report on experiments done on rats and one experiment where people were reading letters:

Other experiments supplied more direct evidence that sleep is crucial for learning. Human subjects were trained to identify letters that appeared for a blink of an eye on a computer screen. Then, half of the subjects were sent home to sleep, while the other half were deprived of sleep for the entire night, and only then went home to rest. Two days later when all the subjects were already rested and refreshed, the scientists checked their ability to read the flashing letters. None of the participants were tired, and yet the people who went to sleep right after the training performed much better than the ones who went to sleep a day later. This suggests that the night sleep immediately after the activity was crucial for gaining the most from the training session. Without it, the training was much less effective.

More importantly, in this study on learning and reducing behavior to some Bayesian process, the experiment reveals that sleep actual helps remember Bayesian unimodal and bimodal priors. An interesting question I have about this study is: at what point people become experts, when they have learned bimodal priors, tri-modal priors ? is the time needed to learn a multi-modal prior (say n), an O(n^p) process where p > 1 ? Does sleep help reduce the complexity from a high p to a low p ?

While we are it, it looks like that a shift from sleeping late to rising early may signal end of adolescence, many people I know in their forties, are still teen-agers...

Static Typing in Python and the Academie Francaise

This thread in comp.lang.python reminds me of the reason why certain languages sometimes just die. For people not seeing the problem, Guido Van Rossum, the maker of Python, is expecting to add an optional feature in the Python language that will allow people to tell the program the type of variable one is using (optional static typing). The point people against the proposal are making is that if this feature were to be implemented it would eventually de facto become a requirement and kill the language altogether. This is not a small point by any means, I switched from Fortran to Matlab in a matter of minutes when I realized I did not need to tell the program that such and such variables were of a certain form/type. The parallel I would suggest using is that of the French language and the French Academy. The French Academy was set up in order to unify the country around a language thereby enabling every citizen to acurately understand the laws of the country. By providing a better framework, it has fulfilled its role of unifying the country around a language at the expense of running out of steam with languages that were a little bit more dynamic (american english). In the end, it's popularity has decreased substantially when spoken outside of France and many people think this is due to the nitpicking aspect of adding new features so that they would fit a larger and consistent framework.

What is surprising to me is that the feature that made me run away from Fortran is likely to make me run away from python if it becomes a requirement. The matlab or Lisp folks sure aren't making this mistake. Could another solution to this whole question be to have very intelligent Python compilers ?

The meaning of life has always been 47

I liked the answer from Roger Schank to the Edge survey question: what do you believe and cannot prove. He answered Irrational choices.
A similar thought crossed my mind about the the DARPA Grand Challenge racers being truely artificial intelligence beings. If they are, then we should not be surprised to see one car making it five meters to the finish line of the race and completely stop while the crowd is yelling at it to cross the line, a situation not unlike that of the irrational choice made by the sprinter in "the loneliless of the long distance runner".

Traveling Interferometry

Here is an interesting application of solving the Traveling Salesman Problem: position the different spacecrafts that will eventually be part of a very large inteferometer. Some related talks can be found here.