TheMrguiller commited on
Commit
180af44
1 Parent(s): 1420410

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +31 -57
README.md CHANGED
@@ -22,6 +22,15 @@ dataset_info:
22
  num_examples: 4242
23
  download_size: 621545899
24
  dataset_size: 684003910.318
 
 
 
 
 
 
 
 
 
25
  ---
26
  # Dataset Card for "ScienceQA"
27
 
@@ -35,84 +44,49 @@ dataset_info:
35
 
36
  ### Dataset Summary
37
 
38
- [More Information Needed]
 
 
39
 
40
  ### Supported Tasks and Leaderboards
41
 
42
- [More Information Needed]
43
 
44
  ### Languages
45
 
46
- [More Information Needed]
47
 
48
  ## Dataset Structure
49
 
50
- ### Data Instances
51
-
52
- [More Information Needed]
53
-
54
  ### Data Fields
55
 
56
- [More Information Needed]
 
 
 
 
 
57
 
58
  ### Data Splits
59
 
60
- [More Information Needed]
61
-
62
- ## Dataset Creation
63
-
64
- ### Curation Rationale
65
-
66
- [More Information Needed]
67
-
68
- ### Source Data
69
-
70
- #### Initial Data Collection and Normalization
71
-
72
- [More Information Needed]
73
-
74
- #### Who are the source language producers?
75
-
76
- [More Information Needed]
77
-
78
- ### Annotations
79
-
80
- #### Annotation process
81
-
82
- [More Information Needed]
83
-
84
- #### Who are the annotators?
85
-
86
- [More Information Needed]
87
-
88
- ### Personal and Sensitive Information
89
-
90
- [More Information Needed]
91
 
92
  ## Considerations for Using the Data
93
 
94
- ### Social Impact of Dataset
95
-
96
- [More Information Needed]
97
-
98
- ### Discussion of Biases
99
-
100
- [More Information Needed]
101
-
102
- ### Other Known Limitations
103
-
104
- [More Information Needed]
105
 
106
  ## Additional Information
107
 
108
  ### Dataset Curators
109
 
110
- [More Information Needed]
111
-
112
- ### Licensing Information
113
-
114
- [More Information Needed]
115
 
116
  ### Citation Information
117
-
118
- [More Information Needed]
 
 
 
 
 
 
 
22
  num_examples: 4242
23
  download_size: 621545899
24
  dataset_size: 684003910.318
25
+ task_categories:
26
+ - question-answering
27
+ - visual-question-answering
28
+ language:
29
+ - en
30
+ tags:
31
+ - code
32
+ size_categories:
33
+ - 100B<n<1T
34
  ---
35
  # Dataset Card for "ScienceQA"
36
 
 
44
 
45
  ### Dataset Summary
46
 
47
+ ScienceQA is collected from elementary and high school science curricula, and contains 21,208 multimodal multiple-choice science questions. Out of the questions in ScienceQA, 10,332 (48.7%) have an image context, 10,220 (48.2%) have a text context, and 6,532 (30.8%) have both. Most questions are annotated with grounded lectures (83.9%) and detailed explanations (90.5%). The lecture and explanation provide general external knowledge and specific reasons, respectively, for arriving at the correct answer. To the best of our knowledge, ScienceQA is the first large-scale multimodal dataset that annotates lectures and explanations for the answers.
48
+
49
+ ScienceQA, in contrast to previous datasets, has richer domain diversity from three subjects: natural science, language science, and social science. Questions in each subject are categorized first by the topic (Biology, Physics, Chemistry, etc.), then by the category (Plants, Cells, Animals, etc.), and finally by the skill (Classify fruits and vegetables as plant parts, Identify countries of Africa, etc.). ScienceQA features 26 topics, 127 categories, and 379 skills that cover a wide range of domains.
50
 
51
  ### Supported Tasks and Leaderboards
52
 
53
+ The dataset is prepared to used it for visual question-answering.
54
 
55
  ### Languages
56
 
57
+ The dataset is in english
58
 
59
  ## Dataset Structure
60
 
 
 
 
 
61
  ### Data Fields
62
 
63
+ - `image`: This field has the image, which is the context given to the model.
64
+ - `question`: This field incorporates the question that has to answer the model from the image context.
65
+ - `choices`: Multiple choice selection.
66
+ - `answer`: The answer from the multiple choice.
67
+ - `solution`: The chain of thought process of the solution selection.
68
+ - `CTH`: A flag that indicates whether it doesnt have chain of thought in that row.
69
 
70
  ### Data Splits
71
 
72
+ The dataset is split in 80% train and 20% test.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
73
 
74
  ## Considerations for Using the Data
75
 
76
+ The dataset is well balanced in order to get really got result when used in multimodal models.
 
 
 
 
 
 
 
 
 
 
77
 
78
  ## Additional Information
79
 
80
  ### Dataset Curators
81
 
82
+ The curators of this dataset where the students from the Masters degree in Computation and Inteligent Systems from University of Deusto.
 
 
 
 
83
 
84
  ### Citation Information
85
+ ```
86
+ @inproceedings{lu2022learn,
87
+ title={Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answering},
88
+ author={Lu, Pan and Mishra, Swaroop and Xia, Tony and Qiu, Liang and Chang, Kai-Wei and Zhu, Song-Chun and Tafjord, Oyvind and Clark, Peter and Ashwin Kalyan},
89
+ booktitle={The 36th Conference on Neural Information Processing Systems (NeurIPS)},
90
+ year={2022}
91
+ }
92
+ ```