Q&A: AI-Based Analysis in the Talmud
AI-Based Analysis in the Talmud
Question
Hello Rabbi Michi,
I’m a computer science student and am currently looking for ideas for a thesis. Because of my familiarity with artificial intelligence technologies and various kinds of big data (among other things, in the context of text analysis), I had the idea of trying to study the Talmudic text using these tools and to extract interesting insights.
The main problem I’ve run into is that it’s hard to “teach” a computer to study a Talmudic passage or analyze an argument, let alone to “understand” the give-and-take of the discussion—almost every AI algorithm requires already-analyzed data in order to train the model on it; to the best of my knowledge, such a database does not exist (an accessible Talmud database systematically divided into passages, arguments, difficulties, resolutions, and the like), so right now I’m looking for research topics that could yield interesting insights without forcing me to manually map out dozens of passages.
Does the Rabbi have any ideas, thoughts, or references to sources that might help me?
Thank you very much in advance!
Answer
It needs some thought. One possibility is to try classifying—supervised, or more interestingly unsupervised—Talmudic passages or parts of passages by structure and style, and to check whether that can confirm or refute the standard classifications in Talmudic scholarship of different types and sources.
Discussion on Answer
Just to note: if all that’s needed to collect data is to manually map “dozens” of passages, then that’s really peanuts. Most likely, anything non-trivial you’ll want to do will require a huge amount of work collecting and organizing the information, whether in code or by hand.
Hey Multivok, I’m just at the beginning of my path in this field. I’d be happy to talk and see what ended up happening with this idea, and whether it’s possible to be part of it.
I don’t understand this stuff at all, but maybe try working on/with the Talmudic Encyclopedia.