{
  "fetched_at": "2026-10-24T23:34:07Z",
  "source_url": "https://www.kaggle.com/competitions/ai-inductions-21-nlp/data",
  "pages": [
    {
      "name": "data-description",
      "content": "About Dataset\nThe training data is a CSV file containing the question, context and the corresponding answer. We have also provided the start and end indexes of the context from which the answer was extracted. Your predictions will have to be made on the test CSV which contains the questions and the context. Your predictions have to be inserted into the PredictionString column in the submission CSV and the submission has to be made via the google drive link.\n\nFile Descriptions\n1. train.csv - A CSV file containing the training data\n2. test.csv - A CSV file containing the test data to make predictions on\n3. submission.csv - A CSV file containing id to which prediction string should be added to make submission.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "Welcome to the AI Club inductions for the year 2021. It’s been a year since the establishment of the club and now it’s time to open our doors to new talent. This year we’ve decided to set up two competions as a part of the induction process, a Natural Language Processing (NLP) task and a Computer Vision (CV) task. Participants can choose to take part in either competitions. We look forward to see your innovative submission. Best of luck and get hacking.\n\nAbout this competition\nThis competition aims to build a simple question answering system. You will have to train an NLP model that takes a question and context as input and outputs the corresponding answer. The leaderboard will show you how your model fares over other participants and this will serve as a major criterion for this induction process.\n\nSubmission\nTo make a submission you will have to add your predictions to the test CSV file(sample csv file is there in \"Data\" section). You will be receiving a mail from AI Club NITC where you will be shared a folder. This csv file has to be uploaded onto the google drive shared folder. Within a span 30 minutes to 1 hour, your position will be updated on the leaderboard, the link for the leaderboard has been given below.\n\nLeaderboard link -https://drive.google.com/drive/folders/1pGt3ROFH0UnV2OK7gMfJLPACwGmtVcYu?usp=sharing\n\nTo save csv file:\ndf.to_csv(\"submission.csv\", index = False)\ndf = DataFrame of the prediction csv file\n\nGuidelines\n1. Participants may participate in only one competition (NLP or CV).\n2. The competition is open to all B19 and B20 students.\n3. Participants are restricted to use only the data that has been provided on the respective competition pages. However, the use of pertained models is allowed.\n4. All submissions must be made via the google drive link, any submission made via Kaggle will not be considered.\n5. All the submission file must be of csv format.\n6.Be careful while uploading the file. Ensure that you upload the correct file into the folder. You will have a maximum of 5 submission chances daily.\n7. Any submission after 11:59 am on 15/10/2021 will not be considered.\n\n\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "Submissions are evaluated on a Jaccard-based FBeta score between predicted texts and ground truth texts, with Beta = 0.5 (a micro F0.5 score).\n\nThe following is Python code for calculating the Jaccard score for a single prediction string against a single ground truth string. Note that the overall score for a sample uses Jaccard to compare multiple ground truth and prediction strings that are pipe-delimited - this code does not handle that process or the final micro F-beta calculation.\n\ndef jaccard(str1, str2): \n    a = set(str1.lower().split()) \n    b = set(str2.lower().split())\n    c = a.intersection(b)\n    return float(len(c)) / (len(a) + len(b) - len(c))\n\nAny matched predictions where the Jaccard score meets or exceeds the threshold of 0.5 are counted as true positives (TP), the remainder as false positives (FP).\nAny unmatched predictions are counted as false positives (FP).\nAny ground truths with no nearest predictions are counted as false negatives (FN).\nAll TP, FP and FN across all samples are used to calculate a final micro F0.5 score. (Note that a micro F score does precisely this, creating one pool of TP, FP and FN that is used to calculate a score for the entire set of predictions.)",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}