7
$\begingroup$

I came across a news article from 2018 where the president of India was saying that Sanskrit is the best language for ML/AI. I have no idea regarding his qualification on either AI or Sanskrit to say this but this idea has been floated earlier in the context of NLP. Specifically, Rick Briggs had said so in 1985.

I know elementary Sanskrit and I know NLP. I can understand the point that as a strongly declined language Sanskrit is less ambiguous than say English as the position of words in a sentence are not important. Add to it the fact that words are also declined (that's not the technical term for verbs and I am not sure what is) so that verb gender and number help identify which entity they refer to.

However, that point was valid in 1985. My question is that post the Deep Learning revolution of the last couple of years is it still relevant? Especially given the fact that humans still have to first learn Sanskrit in case NLP is done in Sanskrit.

Of course, as can be guessed from the tone of the question, I am of the opinion that Sanskrit is not relevant for AI now but I wanted to know if someone who works in AI thinks otherwise and if so what is their reason so think so.

$\endgroup$
3
  • $\begingroup$ In my opinion, the usefulness if Sanskrit has been covered by programming languages. Since I think your question is particularly about knowledge representation in Sanskrit, programming languages can do it in a much better way, since the output of a program is always deterministic irrespective of interpretation. So I don't think it's relevant, although I have seen lots of Indian researchers trying to make language models of Sanskrit. $\endgroup$
    – user9947
    Commented Jan 12, 2020 at 13:38
  • 1
    $\begingroup$ Sorry I don't understand and someone else has mentioned something similar in private communications so I'd be grateful if you could elaborate. Why are we comparing programming languages with Sanskrit for NLP? I thought the idea was that humans communicate in Sanskrit and then an AI system can work on it. If its about programming languages there were plenty of them around in 1985 so what was Briggs talking about? $\endgroup$ Commented Jan 12, 2020 at 19:31
  • $\begingroup$ I think (not sure) the idea you propose is that Sanskrit is an unambiguous language and hence should be used for AI related tasks. Programming languages achieve the same thing. As far as I know NLP is for the computer to make sense of human language and thus it's sole purpose is deciphering of confusing languages like English. $\endgroup$
    – user9947
    Commented Jan 13, 2020 at 9:13

1 Answer 1

2
$\begingroup$

I came across a news article from 2018 where the president of India was saying that Sanskrit is the best language for ML/AI.

A very interesting statement indeed! Globally, there is very little interest in NLP for Sanskrit compared to very many other languages.

Especially given the fact that humans still have to first learn Sanskrit in case NLP is done in Sanskrit.

Most people would say that NLP tools are meant to cater to the needs of humans, instead of requiring humans to learn a new language to be able to benefit from tools (which is a huge barrier).

To answer your question more clearly, it is not that the research community selects languages for research based on "suitability" of the language. Instead, in general the intention of NLP research is to grapple with problems that are relevant, impact many people - or are funded (unfortunately :-).

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .