AI Decrypting Ancient Japanese Scriptures
Do you remember how researchers got their hands on The Egyptian Spellcaster’s book? In case you haven’t heard, here’s the deal. In the year 2014, researchers finally had a breakthrough in deciphering what was called the Handbook of Ritual Powers, a 1300 year old parchment that holds spells for those who wish for love, success, cures to diseases like black jaundice and creepiest of all, exorcism. Can you imagine an Egyptian witch walking through the massive pyramids casting spells and performing witchcraft? That would be a pretty creepy sight!
Not all of these ancient writings are that easy to decipher making researchers pull out all the stops. A 240 page 15 th Century Voynich Manuscript, a handwritten codex, exists to give us the same reality check. This manuscript was written in an alphabet which researchers first claimed to be Hebrew until they weren’t so sure anymore and were back to square one.
Well, so it is now safe for us to blame breakthroughs as such to make a mysterious tingle creep in every time we hear about an ancient scripture being deciphered. The possibility of them hiding many nerve wracking stories from ancient history like the origins of mythical superstitions, insights into war times, magical mysteries or even ancient cures to epidemics might completely change the way we look at the past and even the present world.
Out of many of these undeciphered ancient writings, one has recently come close to the point of closure and the tingle has crept in once again. Kuzushiji, the cursive representation of the Japanese ancient language, Hiragana, that is currently known to less than 0.01 per cent of the world’s population, was used from the 8 th century to the beginning of the 20 th century and the interesting part being, this script was used to write millions of books and documents. But as the era of modernization set in, the usage of these languages had begun to fade and neither commoners and natives nor linguistics could understand and translate this script with no errors. So to fill this gap, researchers are bringing into the picture an artificial intelligence model to make their job easier.
Researcher Tarin Clanuwat and her colleagues are developing a deep learning Optical Character Recognition(OCR) model to transcribe the Kuzushiji scripts. An OCR system is a combination of hardware and software components where the hardware component, such as a specialized circuit board, is used to convert the digital images of physical documents into machine-readable format and the software component processes and transcribes the converted text. It is the software that makes use of deep learning to transcribe this complex script. Labeled datasets from the books of the 17 th century way forward to the 19 th were used to train this model. Despite the complexity of this script, the trained model has an accuracy rate of 85% which is much higher to the prior models.
With artificial intelligence doing the rounds in the transcription business, do you think undeciphered ancient scripts will soon be a thing of the past? We can only wait and find out.
A Computer Science graduate by education and a content writer by profession. Currently fulfilling her zeal to write by putting pen to paper every time she comes across something that is interesting enough to let the world know