Name: ARC: Grade School Science Questions
Creator: Kaggle
License: https://creativecommons.org/publicdomain/zero/1.0/

About this Dataset

ARC: Grade School Science Questions

A Challenge for Advanced Question-Answering Research

By Huggingface Hub [source]

About this dataset

The ARC dataset provides an unparalleled opportunity for ambitious research into all aspects of natural language processing and artificial intelligence. This dataset is a powerful platform to challenge current capacities in question-answering and to uncover fresh discoveries. It contains 7,787 multi-choice questions of grade-school level science, covering a wide range of topics from biology and geology to physics, chemistry, astronomy and environmental science. Particularly noteworthy is the corpus of relevant science sentences that accompanies the dataset. It's divided into an Easy set with only correctly answered questions by our retrieval and co-occurrence algorithms and a Challenge set containing only incorrectly answered questions. With its high quality data design and challenges, this dataset encourages researchers from around the world to explore beyond the boundaries of question answering with neural networks!

More Datasets

For more datasets, click here.

Featured Notebooks

🚨 Your notebook can be here! 🚨!

How to use the dataset

To make use of this data effectively for research purposes or general knowledge growth there are a few steps users should take:

Read through the documentation provided in the README file thoroughly to gain an understanding of the various columns available including question which is simply the text of the question; choices which are the multiple-choice options for each question; and answerKey which is designated letter representing its correct response.

Familiarize yourself with both datasets – ARC Challenge Train Set and ARC Challenge Test Set – including all 7787 multi-choice questions within them using appropriate software tools such as CSV readers/editors such as Microsoft Excel or Google Sheets

Break down each column into individual data points e.g.: separate out questions from answers from choices etc.. This allows users to access specific parts they want better understand certain topics

Utilize data visualization through appropriate software tools like Tableau or PowerBI so that it is easier to interpret data trends

Divide up relevant sections into categories so tests can be conducted more easily and accurately while exploring how well different algorithms perform in recognizing right answers versus wrong ones when presented with similar challenges e.g.: splitting up biology related tests versus physics related tests

Following these steps should enable learning enthusiasts who want engage both casual exploration sessions for knowledge building at an individual level as well as complete large scale study projects across many states/districts whose curriculum could benefit from using this resourceful dataset!

Research Ideas

Developing new methods to answer scientific questions more accurately and quickly using neural networks for natural language processing (NLP).

Training neural networks to recognize the differences between questions that can be answered easily and those that require advanced question-answering techniques.

Using multiple-choice questions to develop interactive educational games that tests students’ science knowledge in a fun environment

Acknowledgements

If you use this dataset in your research, please credit the original authors.
Data Source

License

License: CC0 1.0 Universal (CC0 1.0) - Public Domain Dedication
No Copyright - You can copy, modify, distribute and perform the work, even for commercial purposes, all without asking permission. See Other Information.

Columns

File: ARC-Challenge_test.csv

Column name	Description
question	A grade-school level science question. (String)
choices	Multiple-choice answers for the question. (String)
answerKey	The correct answer for the question. (String)

File: ARC-Easy_test.csv

Column name	Description
question	A grade-school level science question. (String)
choices	Multiple-choice answers for the question. (String)
answerKey	The correct answer for the question. (String)

File: ARC-Challenge_train.csv

Column name	Description
question	A grade-school level science question. (String)
choices	Multiple-choice answers for the question. (String)
answerKey	The correct answer for the question. (String)

Acknowledgements

If you use this dataset in your research, please credit the original authors.
If you use this dataset in your research, please credit Huggingface Hub.

Tables

Arc Challenge Test

@kaggle.thedevastator_arc_grade_school_science_questions.arc_challenge_test

204.68 KB
1172 rows
4 columns


CREATE TABLE arc_challenge_test (
  "id" VARCHAR,
  "question" VARCHAR,
  "choices" VARCHAR,
  "answerkey" VARCHAR
);

Arc Challenge Train

@kaggle.thedevastator_arc_grade_school_science_questions.arc_challenge_train

189.76 KB
1119 rows
4 columns


CREATE TABLE arc_challenge_train (
  "id" VARCHAR,
  "question" VARCHAR,
  "choices" VARCHAR,
  "answerkey" VARCHAR
);

Arc Challenge Validation

@kaggle.thedevastator_arc_grade_school_science_questions.arc_challenge_validation

57.01 KB
299 rows
4 columns


CREATE TABLE arc_challenge_validation (
  "id" VARCHAR,
  "question" VARCHAR,
  "choices" VARCHAR,
  "answerkey" VARCHAR
);

Arc Easy Test

@kaggle.thedevastator_arc_grade_school_science_questions.arc_easy_test

350.8 KB
2376 rows
4 columns


CREATE TABLE arc_easy_test (
  "id" VARCHAR,
  "question" VARCHAR,
  "choices" VARCHAR,
  "answerkey" VARCHAR
);

Arc Easy Train

@kaggle.thedevastator_arc_grade_school_science_questions.arc_easy_train

331.31 KB
2251 rows
4 columns


CREATE TABLE arc_easy_train (
  "id" VARCHAR,
  "question" VARCHAR,
  "choices" VARCHAR,
  "answerkey" VARCHAR
);

Arc Easy Validation

@kaggle.thedevastator_arc_grade_school_science_questions.arc_easy_validation

88.58 KB
570 rows
4 columns


CREATE TABLE arc_easy_validation (
  "id" VARCHAR,
  "question" VARCHAR,
  "choices" VARCHAR,
  "answerkey" VARCHAR
);