compguesswhat / README.md
albertvillanova's picture
Convert dataset to Parquet (#5)
1e853eb verified
metadata
annotations_creators:
  - machine-generated
language_creators:
  - found
language:
  - en
license:
  - unknown
multilinguality:
  - monolingual
size_categories:
  - 100K<n<1M
source_datasets:
  - extended|other-guesswhat
task_categories:
  - visual-question-answering
task_ids:
  - visual-question-answering
paperswithcode_id: compguesswhat
pretty_name: CompGuessWhat?!
dataset_info:
  - config_name: compguesswhat-original
    features:
      - name: id
        dtype: int32
      - name: target_id
        dtype: int32
      - name: timestamp
        dtype: string
      - name: status
        dtype: string
      - name: image
        struct:
          - name: id
            dtype: int32
          - name: file_name
            dtype: string
          - name: flickr_url
            dtype: string
          - name: coco_url
            dtype: string
          - name: height
            dtype: int32
          - name: width
            dtype: int32
          - name: visual_genome
            struct:
              - name: width
                dtype: int32
              - name: height
                dtype: int32
              - name: url
                dtype: string
              - name: coco_id
                dtype: int32
              - name: flickr_id
                dtype: string
              - name: image_id
                dtype: string
      - name: qas
        sequence:
          - name: question
            dtype: string
          - name: answer
            dtype: string
          - name: id
            dtype: int32
      - name: objects
        sequence:
          - name: id
            dtype: int32
          - name: bbox
            sequence: float32
            length: 4
          - name: category
            dtype: string
          - name: area
            dtype: float32
          - name: category_id
            dtype: int32
          - name: segment
            sequence:
              sequence: float32
    splits:
      - name: train
        num_bytes: 123556580
        num_examples: 46341
      - name: validation
        num_bytes: 25441428
        num_examples: 9738
      - name: test
        num_bytes: 25369227
        num_examples: 9621
    download_size: 105349759
    dataset_size: 174367235
  - config_name: compguesswhat-zero_shot
    features:
      - name: id
        dtype: int32
      - name: target_id
        dtype: string
      - name: status
        dtype: string
      - name: image
        struct:
          - name: id
            dtype: int32
          - name: file_name
            dtype: string
          - name: coco_url
            dtype: string
          - name: height
            dtype: int32
          - name: width
            dtype: int32
          - name: license
            dtype: int32
          - name: open_images_id
            dtype: string
          - name: date_captured
            dtype: string
      - name: objects
        sequence:
          - name: id
            dtype: string
          - name: bbox
            sequence: float32
            length: 4
          - name: category
            dtype: string
          - name: area
            dtype: float32
          - name: category_id
            dtype: int32
          - name: IsOccluded
            dtype: int32
          - name: IsTruncated
            dtype: int32
          - name: segment
            sequence:
              - name: MaskPath
                dtype: string
              - name: LabelName
                dtype: string
              - name: BoxID
                dtype: string
              - name: BoxXMin
                dtype: string
              - name: BoxXMax
                dtype: string
              - name: BoxYMin
                dtype: string
              - name: BoxYMax
                dtype: string
              - name: PredictedIoU
                dtype: string
              - name: Clicks
                dtype: string
    splits:
      - name: nd_valid
        num_bytes: 13510589
        num_examples: 5343
      - name: nd_test
        num_bytes: 36228021
        num_examples: 13836
      - name: od_valid
        num_bytes: 14051972
        num_examples: 5372
      - name: od_test
        num_bytes: 32950869
        num_examples: 13300
    download_size: 6548812
    dataset_size: 96741451
configs:
  - config_name: compguesswhat-original
    data_files:
      - split: train
        path: compguesswhat-original/train-*
      - split: validation
        path: compguesswhat-original/validation-*
      - split: test
        path: compguesswhat-original/test-*
  - config_name: compguesswhat-zero_shot
    data_files:
      - split: nd_valid
        path: compguesswhat-zero_shot/nd_valid-*
      - split: nd_test
        path: compguesswhat-zero_shot/nd_test-*
      - split: od_valid
        path: compguesswhat-zero_shot/od_valid-*
      - split: od_test
        path: compguesswhat-zero_shot/od_test-*

Dataset Card for "compguesswhat"

Table of Contents

Dataset Description

Dataset Summary

CompGuessWhat?! is an instance of a multi-task framework for evaluating the quality of learned neural representations, in particular concerning attribute grounding. Use this dataset if you want to use the set of games whose reference scene is an image in VisualGenome. Visit the website for more details: https://compguesswhat.github.io

Supported Tasks and Leaderboards

More Information Needed

Languages

More Information Needed

Dataset Structure

Data Instances

compguesswhat-original

  • Size of downloaded dataset files: 107.21 MB
  • Size of the generated dataset: 174.37 MB
  • Total amount of disk used: 281.57 MB

An example of 'validation' looks as follows.

This example was too long and was cropped:

{
    "id": 2424,
    "image": "{\"coco_url\": \"http://mscoco.org/images/270512\", \"file_name\": \"COCO_train2014_000000270512.jpg\", \"flickr_url\": \"http://farm6.stat...",
    "objects": "{\"area\": [1723.5133056640625, 4838.5361328125, 287.44476318359375, 44918.7109375, 3688.09375, 522.1935424804688], \"bbox\": [[5.61...",
    "qas": {
        "answer": ["Yes", "No", "No", "Yes"],
        "id": [4983, 4996, 5006, 5017],
        "question": ["Is it in the foreground?", "Does it have wings?", "Is it a person?", "Is it a vehicle?"]
    },
    "status": "success",
    "target_id": 1197044,
    "timestamp": "2016-07-08 15:07:38"
}

compguesswhat-zero_shot

  • Size of downloaded dataset files: 4.84 MB
  • Size of the generated dataset: 96.74 MB
  • Total amount of disk used: 101.59 MB

An example of 'nd_valid' looks as follows.

This example was too long and was cropped:

{
    "id": 0,
    "image": {
        "coco_url": "https://s3.amazonaws.com/nocaps/val/004e21eb2e686f40.jpg",
        "date_captured": "2018-11-06 11:04:33",
        "file_name": "004e21eb2e686f40.jpg",
        "height": 1024,
        "id": 6,
        "license": 0,
        "open_images_id": "004e21eb2e686f40",
        "width": 768
    },
    "objects": "{\"IsOccluded\": [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], \"IsTruncated\": [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], \"area\": [3...",
    "status": "incomplete",
    "target_id": "004e21eb2e686f40_30"
}

Data Fields

The data fields are the same among all splits.

compguesswhat-original

  • id: a int32 feature.
  • target_id: a int32 feature.
  • timestamp: a string feature.
  • status: a string feature.
  • id: a int32 feature.
  • file_name: a string feature.
  • flickr_url: a string feature.
  • coco_url: a string feature.
  • height: a int32 feature.
  • width: a int32 feature.
  • width: a int32 feature.
  • height: a int32 feature.
  • url: a string feature.
  • coco_id: a int32 feature.
  • flickr_id: a string feature.
  • image_id: a string feature.
  • qas: a dictionary feature containing:
    • question: a string feature.
    • answer: a string feature.
    • id: a int32 feature.
  • objects: a dictionary feature containing:
    • id: a int32 feature.
    • bbox: a list of float32 features.
    • category: a string feature.
    • area: a float32 feature.
    • category_id: a int32 feature.
    • segment: a dictionary feature containing:
      • feature: a float32 feature.

compguesswhat-zero_shot

  • id: a int32 feature.
  • target_id: a string feature.
  • status: a string feature.
  • id: a int32 feature.
  • file_name: a string feature.
  • coco_url: a string feature.
  • height: a int32 feature.
  • width: a int32 feature.
  • license: a int32 feature.
  • open_images_id: a string feature.
  • date_captured: a string feature.
  • objects: a dictionary feature containing:
    • id: a string feature.
    • bbox: a list of float32 features.
    • category: a string feature.
    • area: a float32 feature.
    • category_id: a int32 feature.
    • IsOccluded: a int32 feature.
    • IsTruncated: a int32 feature.
    • segment: a dictionary feature containing:
      • MaskPath: a string feature.
      • LabelName: a string feature.
      • BoxID: a string feature.
      • BoxXMin: a string feature.
      • BoxXMax: a string feature.
      • BoxYMin: a string feature.
      • BoxYMax: a string feature.
      • PredictedIoU: a string feature.
      • Clicks: a string feature.

Data Splits

compguesswhat-original

train validation test
compguesswhat-original 46341 9738 9621

compguesswhat-zero_shot

nd_valid od_valid nd_test od_test
compguesswhat-zero_shot 5343 5372 13836 13300

Dataset Creation

Curation Rationale

More Information Needed

Source Data

Initial Data Collection and Normalization

More Information Needed

Who are the source language producers?

More Information Needed

Annotations

Annotation process

More Information Needed

Who are the annotators?

More Information Needed

Personal and Sensitive Information

More Information Needed

Considerations for Using the Data

Social Impact of Dataset

More Information Needed

Discussion of Biases

More Information Needed

Other Known Limitations

More Information Needed

Additional Information

Dataset Curators

More Information Needed

Licensing Information

More Information Needed

Citation Information

@inproceedings{suglia-etal-2020-compguesswhat,
    title = "{C}omp{G}uess{W}hat?!: A Multi-task Evaluation Framework for Grounded Language Learning",
    author = "Suglia, Alessandro  and
      Konstas, Ioannis  and
      Vanzo, Andrea  and
      Bastianelli, Emanuele  and
      Elliott, Desmond  and
      Frank, Stella  and
      Lemon, Oliver",
    booktitle = "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics",
    month = jul,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/2020.acl-main.682",
    pages = "7625--7641",
    abstract = "Approaches to Grounded Language Learning are commonly focused on a single task-based final performance measure which may not depend on desirable properties of the learned hidden representations, such as their ability to predict object attributes or generalize to unseen situations. To remedy this, we present GroLLA, an evaluation framework for Grounded Language Learning with Attributes based on three sub-tasks: 1) Goal-oriented evaluation; 2) Object attribute prediction evaluation; and 3) Zero-shot evaluation. We also propose a new dataset CompGuessWhat?! as an instance of this framework for evaluating the quality of learned neural representations, in particular with respect to attribute grounding. To this end, we extend the original GuessWhat?! dataset by including a semantic layer on top of the perceptual one. Specifically, we enrich the VisualGenome scene graphs associated with the GuessWhat?! images with several attributes from resources such as VISA and ImSitu. We then compare several hidden state representations from current state-of-the-art approaches to Grounded Language Learning. By using diagnostic classifiers, we show that current models{'} learned representations are not expressive enough to encode object attributes (average F1 of 44.27). In addition, they do not learn strategies nor representations that are robust enough to perform well when novel scenes or objects are involved in gameplay (zero-shot best accuracy 50.06{\%}).",
}

Contributions

Thanks to @thomwolf, @aleSuglia, @lhoestq for adding this dataset.