amazonmechanicalturk
Latest
Computers can tell when you've been drunk tweeting
In case the rambling string of misspelled words and incoherent thoughts weren't dead giveaways, scientists have developed a method of machine learning to sniff out drunk tweets. Researchers from the University of Rochester collected 11,000 geotagged tweets over a year from two areas: New York and Monroe County, filtering the 140-character notes containing "drunk," "beer," "party" and other libation-related words. From there the school employed Amazon Mechanical Turks to decide whether the person sending the tweets was simply talking about booze or were actually drinking it while tweeting.
Microsoft wants to make you funnier in online chats
Do you constantly struggle to make jokes? Are you one of the millions of folks who wishes they could be spontaneously funny when communicating via the internet? Microsoft Research has some news (PDF) for you. The company's developed a web-based chat system dubbed "Cahoots" that analyzes text in your messages and suggests memes and reaction GIFs based on context and perceived emotion of the sender. It can even auto-generate memes on the fly using existing templates (think QuickMeme). Images were culled from I Can Haz Cheeseburger, Bing and ReactionGIFs.com and at the end of the study, some 738 of Amazon's Mechanical Turks found Cahoots pretty useful. No word's been given about the system actually reaching the public, so in the meantime you might want to start marathoning comedy specials on Netflix to bone up on your timing. [Image credit: See-ming Lee æææ SML/Flickr]
DIY 'Descriptive Camera' captures images, prints out prose
Have you ever wished that cameras could capture not only an image of the scene in front of them, but also describe it to you in plain English? Ok, what if it bypassed the whole "photo" thing and just spit out a slip of paper with words on it? Well, if you're still interested, the impressive Matt Richardson (of Make fame) has a project right up your alley. The Descriptive Camera is a relatively simple device really. A Logitech webcam is connected to a BeagleBone dev board, which is in turn plugged into a thermal printer from Adafruit. Obviously, the tech required to analyze a scene recognize the objects then convert that pile of pixels into a slice prose is outside of the budget and capability of your average DIY project. So Matt turned instead to Mechanical Turk, Amazon's marketplace for HITs or Human Intelligence Tasks. Images captured by the cam are uploaded via the BeagleBone, where an actual person describes what he or she sees and the resulting literature is printed out. For more details and images check out the source.
