0
$\begingroup$

Is there such a thing as Facebook's Dr QA with multiple choice questions? Where the algorithm selects the most likely from 3 possible responses?

The idea is to solve the following problem automatically. Here are 12 questions. Followed by 12*3 answers. The first 3 correspond to the first question, the second three correspond to the second ..

I have a dataset of 2500 multiple choice questions with their answers.

The questions are in Spanish but can be translated automatically. The questions (common knowledge questions) are unrelated and there is no text to find the answers in.

The English translation powered by Google.

[
      "Which of these options does NOT correspond to a traditional sweet?",
      "What autonomous community do the provinces of Huesca, Teruel and Zaragoza form?",
      "What name does the object that is exchanged between the members of the same team in a relay race?"
      "Which of these options does NOT correspond to a megalithic monument?",
      "What do we mean when we talk about the" 2000 effect "?",
      "Which of these options does the term" alamar "refer to?",
      "How many eyelids can we find in the eye of a camel?",
      "In which of these constellations is the star group of The Seven Sisters?"
      "How many cents does it cost to make a euro?"
      "Which of these filmmakers directed the film" Once Upon a Time in America "?",
      "Which of these singers lent his image for a video game?",
      "Until what year was it prohibited in Spain to translate or have a copy of the Bible in Spanish?"
    ],
    [
      "Cachopo",
      "Torrija",
      "Mona de Pascua",
      "Aragon",
      "Catalonia",
      "The Rioja",
      "Dorsal",
      "Taco",
      "Witness",
      "Baldaquino",
      "Dolmen",
      "Menhir",
      "End of the Mayan calendar",
      "Informatical issue",
      "Increase in the hole in the ozone layer",
      "A kind of cuttlefish",
      "An Arab bath",
      "A type of eyelet",
      "one",
      "two",
      "3",
      "Aquarius",
      "Taurus",
      "Sagittarius",
      "3",
      "4,5",
      "5.2",
      "Francis Ford Coppola",
      "Martin Scorsese",
      "Sergio Leone",
      "David Bowie",
      "Freddie Mercury",
      "Mick jagger",
      "1783",
      "1812",
      "1823"
    ],

Here is the set of all questions and answers

$\endgroup$
  • $\begingroup$ Hi, are you only having these 12 questions and having them phrased exactly the same every time? I'm leaning towards this question not being about AI and probably not requiring a NLP if so. $\endgroup$ – Lio Elbammalf Mar 19 at 12:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.