Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror

+ - Using Computer Vision to Predict the Weather->

Submitted by Anonymous Coward
An anonymous reader writes "Getting stuck in the rain one too many times, and sick of weather forecasts that tell me what tomorrow will be like but not whether it'll rain in 15 minutes, I wanted to do something about it. So a friend and I decided to write some software that pulls in open-source NEXRAD radar data from the NOAA, and predicts the weather at your exact location, down to the minute, for the immediate future.

We've got a sexy working demo (check out the video on the linked-to page), and are trying to raise the funds via Kickstarter to bring it to market."

Link to Original Source

Comment: Re:*.R00 (Score 1) 183

by thepotoo (#34862016) Attached to: RapidShare Threatens Suit Over Piracy Allegations

The courts recognize intent. If you're sharing a paragraph of a book with others with the intent to discuss its impact on modern society, that's fine.

If you're sharing the same paragraph with the intent of combining it with other paragraphs hosted by different people, that's copyright infringement.

IANAL

Comment: Re:Non-human intelligences (Score 1) 785

by thepotoo (#34800418) Attached to: Should Dolphins Be Treated As Non-Human Persons?
Octopi are not anywhere near as intelligent as humans or dolphins or even parrots. They are very smart for invertebrates (mostly in the form of hunting techniques), but they simply do not have neurons in sufficient numbers to be considered sentient. Of course, it's possible that they actually are sentient, but if shown, this would overturn well neigh everything we know about consciousness and the brain. Some species of spiders show similar signs of intelligence.

Comment: Re:I'm gonna bite on this one like it's serious. (Score 1) 810

by thepotoo (#34786314) Attached to: Running Your Own Ghost Investigation?

Nice. Finally a good post. This is a fantastic idea: you have a chance to find solid, hard, evidence of something that most people (myself included) believe doesn't exist, and much more importantly, teach people about the scientific method, critical thinking, electromagnetism, and so much more.

There's about two good suggestions here in the comments: The guy talking about infrasound, and the guy who said find an objectively testable prediction. The latter especially is right on the money. What, exactly, constitutes evidence of a ghost? EM? How will you control for cell phones, cameras, faulty wiring, etc.? Temperature? How will you control for drafts? A "feeling" in one particular area? How will you control for infrasound? Include all family/friends in this stage, it's critical that they approve the criteria.

Once you have a list of criteria which suggest the presence of a ghost, establish "control" areas in the house which feature non-supernatural causes to each of these criteria. Keep this a secret from your family/friends. For example, lots of old industrial fans generate infrasound. Set one up behind a door or otherwise out of sight. There's lot of other fun things you can do, too: Grab a sound file from System Shock or Amnesia: Dark Descent and have it play on a hidden speaker system when people are nearby.

Now, the tricky part. Take your family/friends on a bunch of tours around the house. Do this in several small groups, and have each group fill out a quick questionare about the "hauntedness" of each room in the house. Bring along an infrasound detector (someone suggested a microphone, make sure it can record sounds As you've probably guessed, this is an experiment-in-an-experiment: You're testing your family members' willingness to believe in ghosts (hence the surveys), by taking them on a "debunking" expedition. Once you've found everything possible and eliminated it, take them on another tour, this time activating the planted ghost generating equipment - sound effects, infrasound, etc. Make a big deal about not being able to identify the sources. You should probably do this before rather than after with one group, just to better control the experiment.

Anyway, at the end of all this, you'll have tons of data: you can go over, bit by bit, the recordings and make what you will from them (people will of course say that these do not disprove ghosts). But, you can also compare people's surveys on "hauntedness" from both with and without the planted evidence. Since you made a big deal about not being able to find anything the second time around, people should really think the place is haunted. Compare the results of the surveys: BAM: you've just shown that people only believe in ghosts because they can't find a rational explanation for something. Or: BAM: you've just shown that people will still believe in ghosts when they find a rational explanation for something. Win/win for science, doesn't matter whether ghosts are real or not.

This is probably a little more over-the-top than you were looking for, but if you actually go through with it, you could probably get it published in a psychology journal. A much lower budget version would be to randomize people into two groups, one which tries to eliminate the stuff Artifakt listed above and finds causes (even planted "causes"), and another which searches but can't find anything. Compare surveys between the two groups. At any rate, you should report back to us on what you do find (about your friends/family. I think I can predict what you'll find for evidence on ghosts).

Comment: Re:Why Is It Wrong to Call This ESP? (Score 1) 319

by thepotoo (#34779188) Attached to: Journal Article On Precognition Sparks Outrage

T-test is short for Turing test, duh.

In case you're serious, a T-test is a very simple statistical test used when you have two groups of subjects and want to know if there's a statistically significant difference between them. AC was incorrect to say that this article used just a simple T-test, though. It actually uses Stouffer’s Z method, which is a way of combining results from several studies (in this case a bunch of difference sub experiments) to support a single hypothesis. I have never worked with Stouffer’s Z method, so I can't really comment on it's strengths and weaknesses.

The actual experimental design of the experiments was to take a well known psychology paradigm and run it backwards: training someone on a word list after they have already taken the test on it, for example.

Comment: Re:Simple English Wikipedia (Score 2) 266

by thepotoo (#34583410) Attached to: 'Reading Level' Filter Added To Google Search

This is probably correct. PubMed: 98% advanced

Nature: 61% advanced

Science: 94% advanced

PNAS: 99% advanced

Can anyone figure out why science is so much more "advanced" than Nature? Both seem pretty similar to me.

Oh, and by way of a control group:

I Can Has Cheez Burger, surprisingly 11% intermediate

Math

Physicists Say Graphene Could Create Mass 184

Posted by kdawson
from the boson-walks-into-a-church dept.
eldavojohn writes "Graphene has gotten a lot of press lately. The Nobel prize-winning, fastest-spinning, nanobubble-enhanced silicon replacement is theorized to have a new, more outlandish property. As reported by Technology Review's Physics Blog, graphene should be able to create mass inside properly formed nanotubes. According to Abdulaziz Alhaidari's calculations, if one were to roll up graphene into a nanotube, this could compactifiy dimensions (from the sheet's two down to the tube's one), and thus 'the massless equations that describe the behavior of electrons and holes will change to include a term for mass. In effect, compactifying dimensions creates mass.' What once would require a massive high-energy particle accelerator can now be tested with carbon, electricity, and wires, according to the recent paper."
Programming

StarCraft AI Competition Results 113

Posted by Soulskill
from the teaching-skynet-the-zerg-rush dept.
bgweber writes "The StarCraft AI Competition announced last year has come to a conclusion. The competition received 28 bot submissions from universities and teams all over the world. The winner of the competition was UC Berkeley's submission, which executed a novel mutalisk micromanagement strategy. During the conference, a man versus machine exhibition match was held between the top ranking bot and a former World Cyber Games competitor. While the expert player was capable of defeating the best bot, less experienced players were not as successful. Complete results, bot releases, and replays are available at the competition website."

Comment: Re:Open Notes & Well-Designed Exams (Score 1) 870

by thepotoo (#33574376) Attached to: Preventing Networked Gizmo Use During Exams?
If you really need help on an exam due to language related learning differences, you can stop by the access office (most major universities have one). They have people that can read the exam aloud to you, translate it into different languages, or just give you extra time. There's no excuse for giving native speakers an unfair advantage on exams.

If you're not careful, you're going to catch something.

Working...