Thursday, July 22, 2010

Small World - Big Strides

The image above is a tablet found in western Syria in 1920. The symbols on it are of the Ugaritic language -which was last used around 1200 BC.
The discovery was found in 1920, but not deciphered until 1932 - 12 years later.
But that is all about to change.

Now, Scientists are using computer programs to decipher and translate these symbols.
Interestingly enough, these scientists are from MIT and more specifically, Dr. Regina Barzilay (Jon's professor).
The system that she uses repeatedly analyzes the letters (using a computer program) and links the letters and words to map the symbols to the Hebrew language. It took it a matter of hours to do what it previously took 12 years to accomplish.

Previously, decipherment didn't use computers as they weren't thought to be valuable, but according to Dr. Barzilay, "Our aim is to bring to bear the full power of modern machine learning and statistics to this problem".

This process is in the early stages and more and more languages are being "learned" by the computer programs allowing more languages to be appraised.
Aside from learning the lost languages of Etruscan, The Rohonc Codex, Linear A and Vinca Symbols that date back as far as 4000 BC, these programs would prove valuable for google translators as well!

No comments: