Missed TensorFlow Dev Summit? Check out the video playlist. Watch recordings

super_glue

super_glue/boolq (default config)

  • Description:

SuperGLUE (super.gluebenchmark.com/) is a new benchmark styled after GLUE with a new set of more difficult language understanding tasks, improved resources, and a new public leaderboard.

BoolQ (Boolean Questions, Clark et al., 2019a) is a QA task where each example consists of a short passage and a yes/no question about the passage. The questions are provided anonymously and unsolicited by users of the Google search engine, and afterwards paired with a paragraph from a Wikipedia article containing the answer. Following the original work, we evaluate with accuracy.

  • Config description: BoolQ (Boolean Questions, Clark et al., 2019a) is a QA task where each example consists of a short passage and a yes/no question about the passage. The questions are provided anonymously and unsolicited by users of the Google search engine, and afterwards paired with a paragraph from a Wikipedia article containing the answer. Following the original work, we evaluate with accuracy.
  • Homepage: https://github.com/google-research-datasets/boolean-questions
  • Download size: 3.93 MiB
  • Splits:
Split Examples
'test' 3,245
'train' 9,427
'validation' 3,270
  • Features:
FeaturesDict({
    'idx': tf.int32,
    'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=2),
    'passage': Text(shape=(), dtype=tf.string),
    'question': Text(shape=(), dtype=tf.string),
})
  • Citation:
@inproceedings{clark2019boolq,
  title={BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions},
  author={Clark, Christopher and Lee, Kenton and Chang, Ming-Wei, and Kwiatkowski, Tom and Collins, Michael, and Toutanova, Kristina},
  booktitle={NAACL},
  year={2019}
}
@article{wang2019superglue,
  title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
  author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
  journal={arXiv preprint arXiv:1905.00537},
  year={2019}
}

Note that each SuperGLUE dataset has its own citation. Please see the source to
get the correct citation for each contained dataset.

super_glue/cb

  • Description:

SuperGLUE (super.gluebenchmark.com/) is a new benchmark styled after GLUE with a new set of more difficult language understanding tasks, improved resources, and a new public leaderboard.

The CommitmentBank (De Marneffe et al., 2019) is a corpus of short texts in which at least one sentence contains an embedded clause. Each of these embedded clauses is annotated with the degree to which we expect that the person who wrote the text is committed to the truth of the clause. The resulting task framed as three-class textual entailment on examples that are drawn from the Wall Street Journal, fiction from the British National Corpus, and Switchboard. Each example consists of a premise containing an embedded clause and the corresponding hypothesis is the extraction of that clause. We use a subset of the data that had inter-annotator agreement above 0.85. The data is imbalanced (relatively fewer neutral examples), so we evaluate using accuracy and F1, where for multi-class F1 we compute the unweighted average of the F1 per class.

  • Config description: The CommitmentBank (De Marneffe et al., 2019) is a corpus of short texts in which at least one sentence contains an embedded clause. Each of these embedded clauses is annotated with the degree to which we expect that the person who wrote the text is committed to the truth of the clause. The resulting task framed as three-class textual entailment on examples that are drawn from the Wall Street Journal, fiction from the British National Corpus, and Switchboard. Each example consists of a premise containing an embedded clause and the corresponding hypothesis is the extraction of that clause. We use a subset of the data that had inter-annotator agreement above 0.85. The data is imbalanced (relatively fewer neutral examples), so we evaluate using accuracy and F1, where for multi-class F1 we compute the unweighted average of the F1 per class.
  • Homepage: https://github.com/mcdm/CommitmentBank
  • Download size: 73.71 KiB
  • Splits:
Split Examples
'test' 250
'train' 250
'validation' 56
  • Features:
FeaturesDict({
    'hypothesis': Text(shape=(), dtype=tf.string),
    'idx': tf.int32,
    'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=3),
    'premise': Text(shape=(), dtype=tf.string),
})
  • Citation:
@article{de marneff_simons_tonhauser_2019,
  title={The CommitmentBank: Investigating projection in naturally occurring discourse},
  journal={proceedings of Sinn und Bedeutung 23},
  author={De Marneff, Marie-Catherine and Simons, Mandy and Tonhauser, Judith},
  year={2019}
}
@article{wang2019superglue,
  title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
  author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
  journal={arXiv preprint arXiv:1905.00537},
  year={2019}
}

Note that each SuperGLUE dataset has its own citation. Please see the source to
get the correct citation for each contained dataset.

super_glue/copa

  • Description:

SuperGLUE (super.gluebenchmark.com/) is a new benchmark styled after GLUE with a new set of more difficult language understanding tasks, improved resources, and a new public leaderboard.

The Choice Of Plausible Alternatives (COPA, Roemmele et al., 2011) dataset is a causal reasoning task in which a system is given a premise sentence and two possible alternatives. The system must choose the alternative which has the more plausible causal relationship with the premise. The method used for the construction of the alternatives ensures that the task requires causal reasoning to solve. Examples either deal with alternative possible causes or alternative possible effects of the premise sentence, accompanied by a simple question disambiguating between the two instance types for the model. All examples are handcrafted and focus on topics from online blogs and a photography-related encyclopedia. Following the recommendation of the authors, we evaluate using accuracy.

  • Config description: The Choice Of Plausible Alternatives (COPA, Roemmele et al., 2011) dataset is a causal reasoning task in which a system is given a premise sentence and two possible alternatives. The system must choose the alternative which has the more plausible causal relationship with the premise. The method used for the construction of the alternatives ensures that the task requires causal reasoning to solve. Examples either deal with alternative possible causes or alternative possible effects of the premise sentence, accompanied by a simple question disambiguating between the two instance types for the model. All examples are handcrafted and focus on topics from online blogs and a photography-related encyclopedia. Following the recommendation of the authors, we evaluate using accuracy.
  • Homepage: http://people.ict.usc.edu/~gordon/copa.html
  • Download size: 42.96 KiB
  • Splits:
Split Examples
'test' 500
'train' 400
'validation' 100
  • Features:
FeaturesDict({
    'choice1': Text(shape=(), dtype=tf.string),
    'choice2': Text(shape=(), dtype=tf.string),
    'idx': tf.int32,
    'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=2),
    'premise': Text(shape=(), dtype=tf.string),
    'question': Text(shape=(), dtype=tf.string),
})
  • Citation:
@inproceedings{roemmele2011choice,
  title={Choice of plausible alternatives: An evaluation of commonsense causal reasoning},
  author={Roemmele, Melissa and Bejan, Cosmin Adrian and Gordon, Andrew S},
  booktitle={2011 AAAI Spring Symposium Series},
  year={2011}
}
@article{wang2019superglue,
  title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
  author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
  journal={arXiv preprint arXiv:1905.00537},
  year={2019}
}

Note that each SuperGLUE dataset has its own citation. Please see the source to
get the correct citation for each contained dataset.

super_glue/multirc

  • Description:

SuperGLUE (super.gluebenchmark.com/) is a new benchmark styled after GLUE with a new set of more difficult language understanding tasks, improved resources, and a new public leaderboard.

The Multi-Sentence Reading Comprehension dataset (MultiRC, Khashabi et al., 2018) is a true/false question-answering task. Each example consists of a context paragraph, a question about that paragraph, and a list of possible answers to that question which must be labeled as true or false. Question-answering (QA) is a popular problem with many datasets. We use MultiRC because of a number of desirable properties: (i) each question can have multiple possible correct answers, so each question-answer pair must be evaluated independent of other pairs, (ii) the questions are designed such that answering each question requires drawing facts from multiple context sentences, and (iii) the question-answer pair format more closely matches the API of other SuperGLUE tasks than span-based extractive QA does. The paragraphs are drawn from seven domains including news, fiction, and historical text.

  • Config description: The Multi-Sentence Reading Comprehension dataset (MultiRC, Khashabi et al., 2018) is a true/false question-answering task. Each example consists of a context paragraph, a question about that paragraph, and a list of possible answers to that question which must be labeled as true or false. Question-answering (QA) is a popular problem with many datasets. We use MultiRC because of a number of desirable properties: (i) each question can have multiple possible correct answers, so each question-answer pair must be evaluated independent of other pairs, (ii) the questions are designed such that answering each question requires drawing facts from multiple context sentences, and (iii) the question-answer pair format more closely matches the API of other SuperGLUE tasks than span-based extractive QA does. The paragraphs are drawn from seven domains including news, fiction, and historical text.
  • Homepage: https://cogcomp.org/multirc/
  • Download size: 1.06 MiB
  • Splits:
Split Examples
'test' 9,693
'train' 27,243
'validation' 4,848
  • Features:
FeaturesDict({
    'answer': Text(shape=(), dtype=tf.string),
    'idx': FeaturesDict({
        'answer': tf.int32,
        'paragraph': tf.int32,
        'question': tf.int32,
    }),
    'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=2),
    'paragraph': Text(shape=(), dtype=tf.string),
    'question': Text(shape=(), dtype=tf.string),
})
  • Citation:
@inproceedings{MultiRC2018,
    author = {Daniel Khashabi and Snigdha Chaturvedi and Michael Roth and Shyam Upadhyay and Dan Roth},
    title = {Looking Beyond the Surface:A Challenge Set for Reading Comprehension over Multiple Sentences},
    booktitle = {Proceedings of North American Chapter of the Association for Computational Linguistics (NAACL)},
    year = {2018}
}
@article{wang2019superglue,
  title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
  author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
  journal={arXiv preprint arXiv:1905.00537},
  year={2019}
}

Note that each SuperGLUE dataset has its own citation. Please see the source to
get the correct citation for each contained dataset.

super_glue/record

  • Description:

SuperGLUE (super.gluebenchmark.com/) is a new benchmark styled after GLUE with a new set of more difficult language understanding tasks, improved resources, and a new public leaderboard.

(Reading Comprehension with Commonsense Reasoning Dataset, Zhang et al., 2018) is a multiple-choice QA task. Each example consists of a news article and a Cloze-style question about the article in which one entity is masked out. The system must predict the masked out entity from a given list of possible entities in the provided passage, where the same entity may be expressed using multiple different surface forms, all of which are considered correct. Articles are drawn from CNN and Daily Mail. Following the original work, we evaluate with max (over all mentions) token-level F1 and exact match (EM).

  • Config description: (Reading Comprehension with Commonsense Reasoning Dataset, Zhang et al., 2018) is a multiple-choice QA task. Each example consists of a news article and a Cloze-style question about the article in which one entity is masked out. The system must predict the masked out entity from a given list of possible entities in the provided passage, where the same entity may be expressed using multiple different surface forms, all of which are considered correct. Articles are drawn from CNN and Daily Mail. Following the original work, we evaluate with max (over all mentions) token-level F1 and exact match (EM).
  • Homepage: https://sheng-z.github.io/ReCoRD-explorer/
  • Download size: 49.36 MiB
  • Splits:
Split Examples
'test' 10,000
'train' 100,730
'validation' 10,000
  • Features:
FeaturesDict({
    'answers': Sequence(Text(shape=(), dtype=tf.string)),
    'entities': Sequence(Text(shape=(), dtype=tf.string)),
    'idx': FeaturesDict({
        'passage': tf.int32,
        'query': tf.int32,
    }),
    'passage': Text(shape=(), dtype=tf.string),
    'query': Text(shape=(), dtype=tf.string),
})
  • Citation:
@article{zhang2018record,
  title={Record: Bridging the gap between human and machine commonsense reading comprehension},
  author={Zhang, Sheng and Liu, Xiaodong and Liu, Jingjing and Gao, Jianfeng and Duh, Kevin and Van Durme, Benjamin},
  journal={arXiv preprint arXiv:1810.12885},
  year={2018}
}
@article{wang2019superglue,
  title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
  author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
  journal={arXiv preprint arXiv:1905.00537},
  year={2019}
}

Note that each SuperGLUE dataset has its own citation. Please see the source to
get the correct citation for each contained dataset.

super_glue/rte

  • Description:

SuperGLUE (super.gluebenchmark.com/) is a new benchmark styled after GLUE with a new set of more difficult language understanding tasks, improved resources, and a new public leaderboard.

The Recognizing Textual Entailment (RTE) datasets come from a series of annual competitions on textual entailment, the problem of predicting whether a given premise sentence entails a given hypothesis sentence (also known as natural language inference, NLI). RTE was previously included in GLUE, and we use the same data and format as before: We merge data from RTE1 (Dagan et al., 2006), RTE2 (Bar Haim et al., 2006), RTE3 (Giampiccolo et al., 2007), and RTE5 (Bentivogli et al., 2009). All datasets are combined and converted to two-class classification: entailment and not_entailment. Of all the GLUE tasks, RTE was among those that benefited from transfer learning the most, jumping from near random-chance performance (~56%) at the time of GLUE's launch to 85% accuracy (Liu et al., 2019c) at the time of writing. Given the eight point gap with respect to human performance, however, the task is not yet solved by machines, and we expect the remaining gap to be difficult to close.

  • Config description: The Recognizing Textual Entailment (RTE) datasets come from a series of annual competitions on textual entailment, the problem of predicting whether a given premise sentence entails a given hypothesis sentence (also known as natural language inference, NLI). RTE was previously included in GLUE, and we use the same data and format as before: We merge data from RTE1 (Dagan et al., 2006), RTE2 (Bar Haim et al., 2006), RTE3 (Giampiccolo et al., 2007), and RTE5 (Bentivogli et al., 2009). All datasets are combined and converted to two-class classification: entailment and not_entailment. Of all the GLUE tasks, RTE was among those that benefited from transfer learning the most, jumping from near random-chance performance (~56%) at the time of GLUE's launch to 85% accuracy (Liu et al., 2019c) at the time of writing. Given the eight point gap with respect to human performance, however, the task is not yet solved by machines, and we expect the remaining gap to be difficult to close.
  • Homepage: https://aclweb.org/aclwiki/Recognizing_Textual_Entailment
  • Download size: 733.32 KiB
  • Splits:
Split Examples
'test' 3,000
'train' 2,490
'validation' 277
  • Features:
FeaturesDict({
    'hypothesis': Text(shape=(), dtype=tf.string),
    'idx': tf.int32,
    'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=2),
    'premise': Text(shape=(), dtype=tf.string),
})
  • Citation:
@inproceedings{dagan2005pascal,
  title={The PASCAL recognising textual entailment challenge},
  author={Dagan, Ido and Glickman, Oren and Magnini, Bernardo},
  booktitle={Machine Learning Challenges Workshop},
  pages={177--190},
  year={2005},
  organization={Springer}
}
@inproceedings{bar2006second,
  title={The second pascal recognising textual entailment challenge},
  author={Bar-Haim, Roy and Dagan, Ido and Dolan, Bill and Ferro, Lisa and Giampiccolo, Danilo and Magnini, Bernardo and Szpektor, Idan},
  booktitle={Proceedings of the second PASCAL challenges workshop on recognising textual entailment},
  volume={6},
  number={1},
  pages={6--4},
  year={2006},
  organization={Venice}
}
@inproceedings{giampiccolo2007third,
  title={The third pascal recognizing textual entailment challenge},
  author={Giampiccolo, Danilo and Magnini, Bernardo and Dagan, Ido and Dolan, Bill},
  booktitle={Proceedings of the ACL-PASCAL workshop on textual entailment and paraphrasing},
  pages={1--9},
  year={2007},
  organization={Association for Computational Linguistics}
}
@inproceedings{bentivogli2009fifth,
  title={The Fifth PASCAL Recognizing Textual Entailment Challenge.},
  author={Bentivogli, Luisa and Clark, Peter and Dagan, Ido and Giampiccolo, Danilo},
  booktitle={TAC},
  year={2009}
}
@article{wang2019superglue,
  title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
  author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
  journal={arXiv preprint arXiv:1905.00537},
  year={2019}
}

Note that each SuperGLUE dataset has its own citation. Please see the source to
get the correct citation for each contained dataset.

super_glue/wic

  • Description:

SuperGLUE (super.gluebenchmark.com/) is a new benchmark styled after GLUE with a new set of more difficult language understanding tasks, improved resources, and a new public leaderboard.

The Word-in-Context (WiC, Pilehvar and Camacho-Collados, 2019) dataset supports a word sense disambiguation task cast as binary classification over sentence pairs. Given two sentences and a polysemous (sense-ambiguous) word that appears in both sentences, the task is to determine whether the word is used with the same sense in both sentences. Sentences are drawn from WordNet (Miller, 1995), VerbNet (Schuler, 2005), and Wiktionary. We follow the original work and evaluate using accuracy.

  • Config description: The Word-in-Context (WiC, Pilehvar and Camacho-Collados, 2019) dataset supports a word sense disambiguation task cast as binary classification over sentence pairs. Given two sentences and a polysemous (sense-ambiguous) word that appears in both sentences, the task is to determine whether the word is used with the same sense in both sentences. Sentences are drawn from WordNet (Miller, 1995), VerbNet (Schuler, 2005), and Wiktionary. We follow the original work and evaluate using accuracy.
  • Homepage: https://pilehvar.github.io/wic/
  • Download size: 386.93 KiB
  • Splits:
Split Examples
'test' 1,400
'train' 5,428
'validation' 638
  • Features:
FeaturesDict({
    'end1': tf.int32,
    'end2': tf.int32,
    'idx': tf.int32,
    'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=2),
    'sentence1': Text(shape=(), dtype=tf.string),
    'sentence2': Text(shape=(), dtype=tf.string),
    'start1': tf.int32,
    'start2': tf.int32,
    'word': Text(shape=(), dtype=tf.string),
})
  • Citation:
@article{DBLP:journals/corr/abs-1808-09121,
  author={Mohammad Taher Pilehvar and os{'{e} } Camacho{-}Collados},
  title={WiC: 10, 000 Example Pairs for Evaluating Context-Sensitive Representations},
  journal={CoRR},
  volume={abs/1808.09121},
  year={2018},
  url={http://arxiv.org/abs/1808.09121},
  archivePrefix={arXiv},
  eprint={1808.09121},
  timestamp={Mon, 03 Sep 2018 13:36:40 +0200},
  biburl={https://dblp.org/rec/bib/journals/corr/abs-1808-09121},
  bibsource={dblp computer science bibliography, https://dblp.org}
}
@article{wang2019superglue,
  title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
  author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
  journal={arXiv preprint arXiv:1905.00537},
  year={2019}
}

Note that each SuperGLUE dataset has its own citation. Please see the source to
get the correct citation for each contained dataset.

super_glue/wsc

  • Description:

SuperGLUE (super.gluebenchmark.com/) is a new benchmark styled after GLUE with a new set of more difficult language understanding tasks, improved resources, and a new public leaderboard.

The Winograd Schema Challenge (WSC, Levesque et al., 2012) is a reading comprehension task in which a system must read a sentence with a pronoun and select the referent of that pronoun from a list of choices. Given the difficulty of this task and the headroom still left, we have included WSC in SuperGLUE and recast the dataset into its coreference form. The task is cast as a binary classification problem, as opposed to N-multiple choice, in order to isolate the model's ability to understand the coreference links within a sentence as opposed to various other strategies that may come into play in multiple choice conditions. With that in mind, we create a split with 65% negative majority class in the validation set, reflecting the distribution of the hidden test set, and 52% negative class in the training set. The training and validation examples are drawn from the original Winograd Schema dataset (Levesque et al., 2012), as well as those distributed by the affiliated organization Commonsense Reasoning. The test examples are derived from fiction books and have been shared with us by the authors of the original dataset. Previously, a version of WSC recast as NLI as included in GLUE, known as WNLI. No substantial progress was made on WNLI, with many submissions opting to submit only majority class predictions. WNLI was made especially difficult due to an adversarial train/dev split: Premise sentences that appeared in the training set sometimes appeared in the development set with a different hypothesis and a flipped label. If a system memorized the training set without meaningfully generalizing, which was easy due to the small size of the training set, it could perform far below chance on the development set. We remove this adversarial design in the SuperGLUE version of WSC by ensuring that no sentences are shared between the training, validation, and test sets.

However, the validation and test sets come from different domains, with the validation set consisting of ambiguous examples such that changing one non-noun phrase word will change the coreference dependencies in the sentence. The test set consists only of more straightforward examples, with a high number of noun phrases (and thus more choices for the model), but low to no ambiguity.

  • Config description: The Winograd Schema Challenge (WSC, Levesque et al., 2012) is a reading comprehension task in which a system must read a sentence with a pronoun and select the referent of that pronoun from a list of choices. Given the difficulty of this task and the headroom still left, we have included WSC in SuperGLUE and recast the dataset into its coreference form. The task is cast as a binary classification problem, as opposed to N-multiple choice, in order to isolate the model's ability to understand the coreference links within a sentence as opposed to various other strategies that may come into play in multiple choice conditions. With that in mind, we create a split with 65% negative majority class in the validation set, reflecting the distribution of the hidden test set, and 52% negative class in the training set. The training and validation examples are drawn from the original Winograd Schema dataset (Levesque et al., 2012), as well as those distributed by the affiliated organization Commonsense Reasoning. The test examples are derived from fiction books and have been shared with us by the authors of the original dataset. Previously, a version of WSC recast as NLI as included in GLUE, known as WNLI. No substantial progress was made on WNLI, with many submissions opting to submit only majority class predictions. WNLI was made especially difficult due to an adversarial train/dev split: Premise sentences that appeared in the training set sometimes appeared in the development set with a different hypothesis and a flipped label. If a system memorized the training set without meaningfully generalizing, which was easy due to the small size of the training set, it could perform far below chance on the development set. We remove this adversarial design in the SuperGLUE version of WSC by ensuring that no sentences are shared between the training, validation, and test sets.

However, the validation and test sets come from different domains, with the validation set consisting of ambiguous examples such that changing one non-noun phrase word will change the coreference dependencies in the sentence. The test set consists only of more straightforward examples, with a high number of noun phrases (and thus more choices for the model), but low to no ambiguity. * Homepage: https://cs.nyu.edu/faculty/davise/papers/WinogradSchemas/WS.html

  • Download size: 31.98 KiB * Splits:
Split Examples
'test' 146
'train' 554
'validation' 104
  • Features:
FeaturesDict({
    'idx': tf.int32,
    'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=2),
    'span1_index': tf.int32,
    'span1_text': Text(shape=(), dtype=tf.string),
    'span2_index': tf.int32,
    'span2_text': Text(shape=(), dtype=tf.string),
    'text': Text(shape=(), dtype=tf.string),
})
  • Citation:
@inproceedings{levesque2012winograd,
  title={The winograd schema challenge},
  author={Levesque, Hector and Davis, Ernest and Morgenstern, Leora},
  booktitle={Thirteenth International Conference on the Principles of Knowledge Representation and Reasoning},
  year={2012}
}
@article{wang2019superglue,
  title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
  author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
  journal={arXiv preprint arXiv:1905.00537},
  year={2019}
}

Note that each SuperGLUE dataset has its own citation. Please see the source to
get the correct citation for each contained dataset.

super_glue/wsc.fixed

  • Description:

SuperGLUE (super.gluebenchmark.com/) is a new benchmark styled after GLUE with a new set of more difficult language understanding tasks, improved resources, and a new public leaderboard.

The Winograd Schema Challenge (WSC, Levesque et al., 2012) is a reading comprehension task in which a system must read a sentence with a pronoun and select the referent of that pronoun from a list of choices. Given the difficulty of this task and the headroom still left, we have included WSC in SuperGLUE and recast the dataset into its coreference form. The task is cast as a binary classification problem, as opposed to N-multiple choice, in order to isolate the model's ability to understand the coreference links within a sentence as opposed to various other strategies that may come into play in multiple choice conditions. With that in mind, we create a split with 65% negative majority class in the validation set, reflecting the distribution of the hidden test set, and 52% negative class in the training set. The training and validation examples are drawn from the original Winograd Schema dataset (Levesque et al., 2012), as well as those distributed by the affiliated organization Commonsense Reasoning. The test examples are derived from fiction books and have been shared with us by the authors of the original dataset. Previously, a version of WSC recast as NLI as included in GLUE, known as WNLI. No substantial progress was made on WNLI, with many submissions opting to submit only majority class predictions. WNLI was made especially difficult due to an adversarial train/dev split: Premise sentences that appeared in the training set sometimes appeared in the development set with a different hypothesis and a flipped label. If a system memorized the training set without meaningfully generalizing, which was easy due to the small size of the training set, it could perform far below chance on the development set. We remove this adversarial design in the SuperGLUE version of WSC by ensuring that no sentences are shared between the training, validation, and test sets.

However, the validation and test sets come from different domains, with the validation set consisting of ambiguous examples such that changing one non-noun phrase word will change the coreference dependencies in the sentence. The test set consists only of more straightforward examples, with a high number of noun phrases (and thus more choices for the model), but low to no ambiguity.

This version fixes issues where the spans are not actually substrings of the text.

  • Config description: The Winograd Schema Challenge (WSC, Levesque et al., 2012) is a reading comprehension task in which a system must read a sentence with a pronoun and select the referent of that pronoun from a list of choices. Given the difficulty of this task and the headroom still left, we have included WSC in SuperGLUE and recast the dataset into its coreference form. The task is cast as a binary classification problem, as opposed to N-multiple choice, in order to isolate the model's ability to understand the coreference links within a sentence as opposed to various other strategies that may come into play in multiple choice conditions. With that in mind, we create a split with 65% negative majority class in the validation set, reflecting the distribution of the hidden test set, and 52% negative class in the training set. The training and validation examples are drawn from the original Winograd Schema dataset (Levesque et al., 2012), as well as those distributed by the affiliated organization Commonsense Reasoning. The test examples are derived from fiction books and have been shared with us by the authors of the original dataset. Previously, a version of WSC recast as NLI as included in GLUE, known as WNLI. No substantial progress was made on WNLI, with many submissions opting to submit only majority class predictions. WNLI was made especially difficult due to an adversarial train/dev split: Premise sentences that appeared in the training set sometimes appeared in the development set with a different hypothesis and a flipped label. If a system memorized the training set without meaningfully generalizing, which was easy due to the small size of the training set, it could perform far below chance on the development set. We remove this adversarial design in the SuperGLUE version of WSC by ensuring that no sentences are shared between the training, validation, and test sets.

However, the validation and test sets come from different domains, with the validation set consisting of ambiguous examples such that changing one non-noun phrase word will change the coreference dependencies in the sentence. The test set consists only of more straightforward examples, with a high number of noun phrases (and thus more choices for the model), but low to no ambiguity.

This version fixes issues where the spans are not actually substrings of the text. * Homepage: https://cs.nyu.edu/faculty/davise/papers/WinogradSchemas/WS.html

  • Download size: 31.98 KiB * Splits:
Split Examples
'test' 146
'train' 554
'validation' 104
  • Features:
FeaturesDict({
    'idx': tf.int32,
    'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=2),
    'span1_index': tf.int32,
    'span1_text': Text(shape=(), dtype=tf.string),
    'span2_index': tf.int32,
    'span2_text': Text(shape=(), dtype=tf.string),
    'text': Text(shape=(), dtype=tf.string),
})
  • Citation:
@inproceedings{levesque2012winograd,
  title={The winograd schema challenge},
  author={Levesque, Hector and Davis, Ernest and Morgenstern, Leora},
  booktitle={Thirteenth International Conference on the Principles of Knowledge Representation and Reasoning},
  year={2012}
}
@article{wang2019superglue,
  title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
  author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
  journal={arXiv preprint arXiv:1905.00537},
  year={2019}
}

Note that each SuperGLUE dataset has its own citation. Please see the source to
get the correct citation for each contained dataset.

super_glue/axb

  • Description:

SuperGLUE (super.gluebenchmark.com/) is a new benchmark styled after GLUE with a new set of more difficult language understanding tasks, improved resources, and a new public leaderboard.

An expert-constructed, diagnostic dataset that automatically tests models for a broad range of linguistic, commonsense, and world knowledge. Each example in this broad-coverage diagnostic is a sentence pair labeled with a three-way entailment relation (entailment, neutral, or contradiction) and tagged with labels that indicate the phenomena that characterize the relationship between the two sentences. Submissions to the GLUE leaderboard are required to include predictions from the submission's MultiNLI classifier on the diagnostic dataset, and analyses of the results were shown alongside the main leaderboard. Since this broad-coverage diagnostic task has proved difficult for top models, we retain it in SuperGLUE. However, since MultiNLI is not part of SuperGLUE, we collapse contradiction and neutral into a single not_entailment label, and request that submissions include predictions on the resulting set from the model used for the RTE task.

  • Config description: An expert-constructed, diagnostic dataset that automatically tests models for a broad range of linguistic, commonsense, and world knowledge. Each example in this broad-coverage diagnostic is a sentence pair labeled with a three-way entailment relation (entailment, neutral, or contradiction) and tagged with labels that indicate the phenomena that characterize the relationship between the two sentences. Submissions to the GLUE leaderboard are required to include predictions from the submission's MultiNLI classifier on the diagnostic dataset, and analyses of the results were shown alongside the main leaderboard. Since this broad-coverage diagnostic task has proved difficult for top models, we retain it in SuperGLUE. However, since MultiNLI is not part of SuperGLUE, we collapse contradiction and neutral into a single not_entailment label, and request that submissions include predictions on the resulting set from the model used for the RTE task.
  • Homepage: https://gluebenchmark.com/diagnostics
  • Download size: 33.15 KiB
  • Splits:
Split Examples
'test' 1,104
  • Features:
FeaturesDict({
    'idx': tf.int32,
    'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=2),
    'sentence1': Text(shape=(), dtype=tf.string),
    'sentence2': Text(shape=(), dtype=tf.string),
})
  • Citation:
@article{wang2019superglue,
  title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
  author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
  journal={arXiv preprint arXiv:1905.00537},
  year={2019}
}

Note that each SuperGLUE dataset has its own citation. Please see the source to
get the correct citation for each contained dataset.

super_glue/axg

  • Description:

SuperGLUE (super.gluebenchmark.com/) is a new benchmark styled after GLUE with a new set of more difficult language understanding tasks, improved resources, and a new public leaderboard.

Winogender is designed to measure gender bias in coreference resolution systems. We use the Diverse Natural Language Inference Collection (DNC; Poliak et al., 2018) version that casts Winogender as a textual entailment task. Each example consists of a premise sentence with a male or female pronoun and a hypothesis giving a possible antecedent of the pronoun. Examples occur in minimal pairs, where the only difference between an example and its pair is the gender of the pronoun in the premise. Performance on Winogender is measured with both accuracy and the gender parity score: the percentage of minimal pairs for which the predictions are the same. We note that a system can trivially obtain a perfect gender parity score by guessing the same class for all examples, so a high gender parity score is meaningless unless accompanied by high accuracy. As a diagnostic test of gender bias, we view the schemas as having high positive predictive value and low negative predictive value; that is, they may demonstrate the presence of gender bias in a system, but not prove its absence.

  • Config description: Winogender is designed to measure gender bias in coreference resolution systems. We use the Diverse Natural Language Inference Collection (DNC; Poliak et al., 2018) version that casts Winogender as a textual entailment task. Each example consists of a premise sentence with a male or female pronoun and a hypothesis giving a possible antecedent of the pronoun. Examples occur in minimal pairs, where the only difference between an example and its pair is the gender of the pronoun in the premise. Performance on Winogender is measured with both accuracy and the gender parity score: the percentage of minimal pairs for which the predictions are the same. We note that a system can trivially obtain a perfect gender parity score by guessing the same class for all examples, so a high gender parity score is meaningless unless accompanied by high accuracy. As a diagnostic test of gender bias, we view the schemas as having high positive predictive value and low negative predictive value; that is, they may demonstrate the presence of gender bias in a system, but not prove its absence.
  • Homepage: https://github.com/rudinger/winogender-schemas
  • Download size: 10.17 KiB
  • Splits:
Split Examples
'test' 356
  • Features:
FeaturesDict({
    'hypothesis': Text(shape=(), dtype=tf.string),
    'idx': tf.int32,
    'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=2),
    'premise': Text(shape=(), dtype=tf.string),
})
  • Citation:
@inproceedings{rudinger-EtAl:2018:N18,
  author    = {Rudinger, Rachel  and  Naradowsky, Jason  and  Leonard, Brian  and  {Van Durme}, Benjamin},
  title     = {Gender Bias in Coreference Resolution},
  booktitle = {Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies},
  month     = {June},
  year      = {2018},
  address   = {New Orleans, Louisiana},
  publisher = {Association for Computational Linguistics}
}

@article{wang2019superglue,
  title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
  author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
  journal={arXiv preprint arXiv:1905.00537},
  year={2019}
}

Note that each SuperGLUE dataset has its own citation. Please see the source to
get the correct citation for each contained dataset.