Share This Author
Challenges in Automated Debiasing for Toxic Language Detection
- Xuhui Zhou, Maarten Sap, Swabha Swayamdipta, Noah A. Smith, Yejin Choi
- Computer ScienceEACL
- 29 January 2021
The findings show that debiasing a model trained on biased toxic language data is not as effective as simply relabeling the data to remove existing biases, and proposes an automatic, dialect-aware data correction method, as a proof-of-concept.
Evaluating Commonsense in Pre-trained Language Models
This work studies the commonsense ability of GPT, BERT, XLNet, and RoBERTa by testing them on seven challenging benchmarks, finding that language modeling and its variants are effective objectives for promoting models' commonsens ability while bi-directional context and larger training set are bonuses.
Multilevel Text Alignment with Cross-Document Attention
This work proposes a new learning approach that equips previously established hierarchical attention encoders for representing documents with a cross-document attention component, enabling structural comparisons across different levels (document-to-document and sentence- to-document).
Annotators with Attitudes: How Annotator Beliefs And Identities Bias Toxic Language Detection
- Maarten Sap, Swabha Swayamdipta, Laura Vianna, Xuhui Zhou, Yejin Choi, Noah A. Smith
- 15 November 2021
This work disentangle what is annotated as toxic by considering posts with three characteristics: anti-Black language, African American English dialect, and vulgarity, and shows strong associations between annotator identity and beliefs and their ratings of toxicity.
Linguistically-Informed Transformations (LIT): A Method for Automatically Generating Contrast Sets
- Chuanrong Li, Lin Shengshuo, L. Liu, Xinyi Wu, Xuhui Zhou, Shane Steinert-Threlkeld
- Computer ScienceBLACKBOXNLP
- 16 October 2020
This work proposes a Linguistically-Informed Transformation (LIT) method to automatically generate contrast sets, which enables practitioners to explore linguistic phenomena of interests as well as compose different phenomena.
Learning to translate by learning to communicate
Two variants of EC Fine-Tuning are presented, one of which outperforms a backtranslation-based baseline in 6/8 translation settings, and proves especially beneﬁcial for the very low-resource languages of Nepali and Sinhala.
RPD: A Distance Function Between Word Embeddings
This paper proposes a novel metric called Relative Pairwise Inner Product Distance (RPD) to quantify the distance between different sets of word embeddings and investigates the influence of different training processes and corpora to shed light on the poorly understood word embedDings.