Open Source Homeland

The people over at  reCAPTCHA? have a brilliant  idea. They seek to solve two problems at once:

• How to authenticate people who sign in to websites
• How to digitize millions of pages of books

The first problem is already solved with, among other things, CAPTCHA— those hard-to-read nonsense words we’ve all come across. The second problem can be solved with the first solutions— why not use real words, taken from real books, and round-trip those word back into the digitized books so they’re available in digital form? It takes the Mechanical Turk concept even farther.

About 60 million CAPTCHAs are solved by humans around the world every day. In each case, roughly ten seconds of human time are being spent. Individually, thats not a lot of time, but in aggregate these little puzzles consume more than 150,000 hours of work each day. What if we could make positive use of this human effort? reCAPTCHA does exactly that by channeling the effort spent solving CAPTCHAs online into “reading” books.

Similar to the Who Is Sick? database, this Mechanical turk HIT seeks to collect information about disease from the crowd.

Find description of a disease outbreak in China in Chinese

We assume you read one or more Chinese newspapers daily.

If you see any description of an infectious disease outbreak or even individual new cases in China, send an email in English describing in 1-3 sentences what the article says and include a hyperlink to the original article.

Interesting activities on Amazon Turk, as seen in this email I got from Amazon. It’s easy to image the utility of Amazon Turk in the analysis of satellite imagery of possible terrorist hideouts (”review these images and indicate if there is a truck in the picture”). Adding a prediction market element, you can show photos of houses and ask “does a terrorist live here?”.

A number of you have asked us to make it easier for you to be kept up to date on Amazon Mechanical Turk. This email is a way for us to share with you what’s happening.Currently, there is a large selection of HITs available on the site, and the mix of tasks changes frequently. Here is a small sample, but we encourage you to visit the site and see what else is available.

Information Locators — Identify and transcribe information from digital documents for Information Locators. To complete these HITs, find and transcribe the requested data from online digital documents.

Dimensional Analytics — Describe the contents of a website for Dimensional Analytics. These HITs ask you to provide the top 3 phrases that best describe a website.

Geospatial Vision — Identify and mark items in photographs such as road signs, lamp posts, storm drains and other features for Geospatial Vision using their custom Flash application.

Collectorz.com — Write a short plot description for movies you have watched. Collectorz.com offers several movie titles that you can choose from to write your own plot description.

Trivia Research Department — Create and edit trivia questions for Trivia Research. To complete these HITs, reseach, write and edit trivia questions in different categories using your own knowledge and other resources.

Amazon UnSpun — Express your opinion by ranking your top three choices in various list categories. This information is used by Amazon’s UnSpun website to populate its community opinion lists on a wide variety of topics.

CastingWords — Transcribe audio content and edit transcripts . CastingWords’ HITs ask you to complete such diverse tasks as transcribing audio content into text, editing transcripts for accuracy and verifying transcription quality.

There’s a lot of activity on Mechanical Turk, so the list of available HITs changes frequently. The best way to stay up to date is to check the site when you have a free minute.