Building a Large Annotated Corpus of Learner English: The NUS Corpus of Learner English

By Daniel Dahlmeier, Hwee Tou Ng, Siew Mei Wu. 2013

402 citations recorded by the source corpus.

Subfields: Natural Language Processing, Evaluation & Benchmarks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Natural Language Processing

In Evaluation & Benchmarks

Citations in the corpus

Instantaneous Grammatical Error Correction with Shallow Aggressive Decoding (2021, methodology): “…in the restricted training setting of BEA-2019 GEC shared task (Bryant et al., 2019): We use Lang-8 Corpus of Learner English (Mizumoto et al., 2011), NUCLE (Dahlmeier et al., 2013), FCE (Yannakoudakis et al., 2011) and W&I+LOCNESS (Granger; Bryant et al., 2019) as our GEC training data.”
A Comparative Study of Synthetic Data Generation Methods for Grammatical Error Correction (2020, background): “, 2012), and the NUCLE corpus from the CoNLL-2014 shared task (Dahlmeier et al., 2013) (both treated as out-of-domain for the two datasets).”
Adapting Sequence Models for Sentence Correction (2017, methodology): “The Lang-8 dataset of corrections is large4 but is crowd-sourced5 and is thus of a different nature than the professionally annotated AESW and NUCLE datasets.”
Lossless Acceleration for Seq2seq Generation with Aggressive Decoding (2022, methodology): “…in the restricted training setting of BEA-2019 GEC shared task (Bryant et al., 2019): We use Lang-8 Corpus of Learner English (Mizumoto et al., 2011), NUCLE (Dahlmeier et al., 2013), FCE (Yannakoudakis et al., 2011) and W&I+LOCNESS (Granger; Bryant et al., 2019) as our GEC training data.”
A Neural Grammatical Error Correction System Built On Better Pre-training and Sequential Transfer Learning (2019, methodology): “3) dataset (Dahlmeier et al., 2013), and the newly released Write & Improve and LOCNESS (W&I+L) datasets (Yannakoudakis et al., 2018).6 For the “Low Resource Track” (§7.6), we use a 3:1 traintest random split of the W&I+L development set, keeping the proportions of proficiency levels the same.”
The BEA-2019 Shared Task on Grammatical Error Correction (2019, background): “NUCLE The National University of Singapore Corpus of Learner English (NUCLE) consists of 1,400 essays written by mainly Asian undergraduate students at the National University of Singapore (Dahlmeier et al., 2013).”
Fluency Boost Learning and Inference for Neural Grammatical Error Correction (2018, methodology): “, 2012), Cambridge Learner Corpus (CLC) (Nicholls, 2003) and NUS Corpus of Learner English (NUCLE) (Dahlmeier et al., 2013) as our original error-corrected training data.”
Quantifying the Evaluation of Heuristic Methods for Textual Data Augmentation (2020, background): “Unlike the sentiment analysis domain, the set of existing resources for the verbosity detection problem is much more limited: NUCLE covers grammatical redundancy (Dahlmeier et al., 2013), and Kashefi et al.”
A Multilayer Convolutional Encoder-Decoder Neural Network for Grammatical Error Correction (2018, methodology): “They trained their models on the publicly available NUCLE (Dahlmeier, Ng, and Wu 2013) and Lang-8 corpora, along with synthesized examples for frequent error types.”
The CoNLL-2013 Shared Task on Grammatical Error Correction (2013, methodology): “All instances of grammatical errors are annotated in NUCLE, and the errors are classified into 27 error types (Dahlmeier et al., 2013).”
Erroneous data generation for Grammatical Error Correction (2019, methodology): “, 2012), NUS Corpus of Learner English (NUCLE) (Dahlmeier et al., 2013), FCE (Yannakoudakis et al.”
Controllable data synthesis method for grammatical error correction (2019, methodology): “All of these studies are based on two public corpora, the largest parallel corpus Lang-8 (Mizumoto et al., 2011) and NUS Corpus of Learner English (NUCLE) (Dahlmeier et al., 2013).”
A Comprehensive Survey of Grammar Error Correction (2020, methodology): “We first introduce several datasets that are most widely used to develop GEC systems based on supervised approach, including NUCLE [3], Lang-8 [4], FCE [5], J FLEG [6], Write&Improve+LOCNESS [7].”
Neural Grammatical Error Correction Systems with Unsupervised Pre-training on Synthetic Data (2019, methodology): “, 2011), NUCLE (Dahlmeier et al., 2013), Lang-8 Corpus of Learner English (Mizumoto et al.”
Artificial error generation for translation-based grammatical error correction (2016, background): “The National University of Singapore Corpus of Learner English (NUCLE) comprises 1,414 essays written in English by non-native undergraduate students at the National University of Singapore (Dahlmeier et al., 2013).”
GECToR – Grammatical Error Correction: Tag, Not Rewrite (2020, methodology): “We use the following datasets for ﬁne-tuning stages II and III: National University of Singapore Corpus of Learner English (NU-CLE) 4 (Dahlmeier et al., 2013), Lang-8 Corpus of Learner English (Lang-8) 5 (Tajiri et al., 2012), FCE dataset 6 (Yannakoudakis et al., 2011), the publicly available part…”
A Comprehensive Survey of Grammatical Error Correction (2021, methodology): “Unlike NUCLE and FCE, annotation of JFLEG involves not only grammatical error correction but also sentence-level rewrite to make source sentence sound more fluent.”
Sequence-to-sequence Pre-training with Data Augmentation for Sentence Rewriting (2019, methodology): “We follow the restricted setting where only the public resources can be used and use public Lang-8 (Mizumoto et al., 2011; Tajiri et al., 2012) and NUCLE (Dahlmeier et al., 2013) dataset as gold parallel data, as most previous work did.”
JFLEG: A Fluency Corpus and Benchmark for Grammatical Error Correction (2017, methodology): “The NUS Corpus of Learner English (NUCLE) contains essays written by students at the National University of Singapore, corrected by two annotators using 27 error codes (Dahlmeier et al., 2013).”
Reaching Human-level Performance in Automatic Grammatical Error Correction: An Empirical Study (2018, methodology): “, 2012), Cambridge Learner Corpus (CLC) (Nicholls, 2003) and NUS Corpus of Learner English (NUCLE) (Dahlmeier et al., 2013) as our original error-corrected training data.”