A rendezvous with a SMT and Stars

Let me assure you in the beginning that,

I am not going to talk technical gibberish in this post.

If I say 'Diwali is the first that I'm celebrating outside his home' is makes some sense. 'This Diwali is the first that I'm celebrating outside his home' makes more sense. But, when I say that it is the translation of 'यह पहली दिवाली है जो मैं अपने घर से बाहर मना रहा हूँ।' then you would say that I have lost my mind.

So, let me introduce you to my partner and translator for the evening, Google Translate. I had been using it for quite some time. It is a really nice service and came to my rescue quite often, like when I tried to register a GMX email account.

Google Translate gave me a very sweet and nice translation of my last post on Deepawali. Have a look at the translation here, the original post is here. To not strain the translator let us not look at my poor poetry.

The first line of the post was a perfect translation.

यह दीपावली मेरे लिए नई है।

This Diwali is new for me.


Had it been the only criteria for judgment, then Google Translate would have been the perfect partner. But, as it is, it is not the only line it translated. The second line brings in some trouble, Google messed up with the person in this sentence. With the first half being in First Person, Google shifted to Third Person in the second half. It translated 'यह पहली दिवाली है जो मैं अपने घर से बाहर मना रहा हूँ।' to 'Diwali is the first that I'm celebrating outside his home'.

Rest of the translation was not a complete joke. Google Translate threw up marvels like 'unity in diversity' for 'अनेकता में भी एकता'. And, it is not the only one, 'festival of victory of good over evil' for 'बुराई पर अच्छाई के जीत का त्यौहार' is another example.

After, looking at the translation I realized how complex my mother tongue Hindi is. I am amazed at the fluency at which I(we) can understand it. Thank God I am not a computer. But, Google Translate is one! And, the chances that computer makes a mistake is almost zero. So, I was abashed when I read Diwali as the '
festival of lights and sanitation' appear in the translation. The moment I hovered my mouse on the text I got the reason, it was the literal translation of 'स्वच्छता व प्रकाश का पर्व'. Technically I can't blame Google. But, I would have liked 'festival of lights and purity' much better.

A little more exploration told me that Google does take context into account, it translated 'शुभ दीपावली' into 'Good illumination'. Now, where does that come from? It had been writing 'दीपावली' as 'Diwali'.

When I delved further I found out that Google Translate is based on STM or Statistical Machine Translation, a machine translation paradigm which was conceived as early as 1949.

I decided not research further and came back to Google Translate to experiment. Next I tried, English to Hindi translated search. Here, we can key in some words in a language and Google translates them in the language of our choice(the language has to be on the list) and searches for pages containing the translation. Google does not stop here but also displays the the translation of the search results in the original language keyed in.

Now, this was going to be fun. I fired up the search with keywords 'Amrita Rao' in English to be translated and searched in Hindi. Here are the search results.

Wow! The translation was perfect, and 1,44,000 Hindi search results! I should not have been surprised, after all she is a big star and Hindi is picking up space on the web. But, there were a few bloopers like this one.







Now, vanity started to creep in and I tried my name but Google could't even translate it. Poor me!!





Incidentally, what surprised me was that the search about Amrita Rao had relevant items even on page 19, but when I keyed in Shahrukh Khan then, though the number of results were more but, the second result on page one itself was out of context. It was not about Shahrukh Khan at all.






Moreover the first result showed that the web is not too friendly towards him.






Since, I will write only sweet truths here I wont discuss this any further on my blog.


So, my evening ended with writing this post and the idea for a new post.

0 comments:

Copyright © - Mayank Abhishek | Sweet & Pure - is proudly powered by Blogger | Template Design Credits- Design Disease