Nanyun Peng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Nanyun Peng has a total of 95 co-authors. Here are the top co-authors by number of shared publications:
- Kai-Wei Chang
Department of Computer Science, University of California Los Angeles, Los Angeles, USA
2 shared publications - Xiang Ren
University of Southern California
2 shared publications - Abhay Zala
UNC Chapel Hill, Chapel Hill, North Carolina, USA
1 shared publication - Adam Badawy
University of Southern California, USA
1 shared publication Alexander Spangher
1 shared publication- Ansel Blume
UIUC, Urbana, IL, USA
1 shared publication
Latest Publications
- Open Access
- 22 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 1387-1410
CoLyricist: Enhancing Lyric Writing with AI through Workflow-Aligned Support
- 22 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 1387-1410
Open AccessCoLyricist: Enhancing Lyric Writing with AI through Workflow-Aligned Support
- Open Access
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 3781-3789
New Job, New Gender? Measuring the Social Bias in Image Generation Models
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 3781-3789
Open AccessNew Job, New Gender? Measuring the Social Bias in Image Generation Models
- Open Access
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 11252-11254
MIRACLE: An Online, Explainable Multimodal Interactive Concept Learning System
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 11252-11254
Open AccessMIRACLE: An Online, Explainable Multimodal Interactive Concept Learning System
- Open Access
Mitigating Bias for Question Answering Models by Tracking Bias Influence
Models of various NLP tasks have been shown to exhibit stereotypes, and the bias in the question answering (QA) models is especially harmful as the output answers might be directly consumed by the end users. There have been datasets to evaluate bias in QA models, while bias mitigation technique for the QA models is still under-explored. In this work, we propose BMBI, an approach to mitigate the bias of multiple-choice QA models. Based on the intuition that a model would lean to be more biased if it learns from a biased example, we measure the bias level of a query instance by observing its influence on another instance. If the influenced instance is more biased, we derive that the query instance is biased. We then use the bias level detected as an optimization objective to form a multi-task learning setting in addition to the original QA task. We further introduce a new bias evaluation metric to quantify bias in a comprehensive and sensitive way. We show that our method could be applied to multiple QA formulations across multiple bias categories. It can significantly reduce the bias level in all 9 bias categories in the BBQ dataset while maintaining comparable QA accuracy.
Open AccessMitigating Bias for Question Answering Models by Tracking Bias Influence
Models of various NLP tasks have been shown to exhibit stereotypes, and the bias in the question answering (QA) models is especially harmful as the output answers might be directly consumed by the end users. There have been datasets to evaluate bias in QA models, while bias mitigation technique for the QA models is still under-explored. In this work, we propose BMBI, an approach to mitigate the bias of multiple-choice QA models. Based on the intuition that a model would lean to be more biased if it learns from a biased example, we measure the bias level of a query instance by observing its influence on another instance. If the influenced instance is more biased, we derive that the query instance is biased. We then use the bias level detected as an optimization objective to form a multi-task learning setting in addition to the original QA task. We further introduce a new bias evaluation metric to quantify bias in a comprehensive and sensitive way. We show that our method could be applied to multiple QA formulations across multiple bias categories. It can significantly reduce the bias level in all 9 bias categories in the BBQ dataset while maintaining comparable QA accuracy.
- Open Access
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 5899-5900
The Second Workshop on Knowledge-Augmented Methods for Natural Language Processing
Language models are being developed and deployed in many applications, "small"-scale and large-scale, generic and specialized, text-only and multimodal, etc. Meanwhile, the missingness of important knowledge causes limitations and safety challenges. The knowledge includes commonsense, world facts, domain expertise, personalization, and especially the unique patterns that need to be discovered from big data applications. Training and inference processes of the language models can be and should be augmented with the knowledge. The first KnowledgeNLP at AAAI 2023 attracted scientists on knowledge augmentation methods towards higher language intelligence. This workshop offers a broad platform to share ideas and discuss various topics, such as (1) synergy between knowledge and language model, (2) scalable architectures that integrate NLP, knowledge graph, and graph learning technologies, (3) KnowledgeNLP for e-commerce, education, and healthcare, (4) human factors and social good in KnowledgeNLP.
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 5899-5900
Open AccessThe Second Workshop on Knowledge-Augmented Methods for Natural Language Processing
Language models are being developed and deployed in many applications, "small"-scale and large-scale, generic and specialized, text-only and multimodal, etc. Meanwhile, the missingness of important knowledge causes limitations and safety challenges. The knowledge includes commonsense, world facts, domain expertise, personalization, and especially the unique patterns that need to be discovered from big data applications. Training and inference processes of the language models can be and should be augmented with the knowledge. The first KnowledgeNLP at AAAI 2023 attracted scientists on knowledge augmentation methods towards higher language intelligence. This workshop offers a broad platform to share ideas and discuss various topics, such as (1) synergy between knowledge and language model, (2) scalable architectures that integrate NLP, knowledge graph, and graph learning technologies, (3) KnowledgeNLP for e-commerce, education, and healthcare, (4) human factors and social good in KnowledgeNLP.
- Open Access
Helpfulness and Fairness of Task-Oriented Dialogue Systems
Goal-oriented dialogue systems aim to help users achieve certain goals. Therefore, how humans perceive their helpfulness is important. However, neither the human-perceived helpfulness of goal-oriented dialogue systems nor its fairness implication has been well studied. In this paper, we study computational measurements of helpfulness. We first formally define a dialogue response as helpful if it is relevant & coherent, useful, and informative to a query. Then, we collect human annotations for the helpfulness of dialogue responses based on our definition and build a classifier to automatically determine the helpfulness of a response. We further propose to use the helpfulness level of a dialogue system towards different user queries to measure the fairness of a dialogue system. Experiments with state-of-the-art dialogue systems under three information-seeking scenarios reveal that existing systems tend to be more helpful for questions regarding concepts from highly-developed countries than less-developed countries, uncovering potential fairness concerns underlying the current goal-oriented dialogue systems.
Open AccessHelpfulness and Fairness of Task-Oriented Dialogue Systems
Goal-oriented dialogue systems aim to help users achieve certain goals. Therefore, how humans perceive their helpfulness is important. However, neither the human-perceived helpfulness of goal-oriented dialogue systems nor its fairness implication has been well studied. In this paper, we study computational measurements of helpfulness. We first formally define a dialogue response as helpful if it is relevant & coherent, useful, and informative to a query. Then, we collect human annotations for the helpfulness of dialogue responses based on our definition and build a classifier to automatically determine the helpfulness of a response. We further propose to use the helpfulness level of a dialogue system towards different user queries to measure the fairness of a dialogue system. Experiments with state-of-the-art dialogue systems under three information-seeking scenarios reveal that existing systems tend to be more helpful for questions regarding concepts from highly-developed countries than less-developed countries, uncovering potential fairness concerns underlying the current goal-oriented dialogue systems.
- Open Access
- 30 April 2023
- Published by Association for Computing Machinery (ACM)
- p. 1358-1364
Detecting Social Media Manipulation in Low-Resource Languages
Social media have been deliberately used for malicious purposes, including political manipulation and disinformation. Most research focuses on high-resource languages. However, malicious actors share content across countries and languages, including low-resource ones. Here, we investigate whether and to what extent malicious actors can be detected in low-resource language settings. We discovered that a high number of accounts posting in Tagalog were suspended as part of Twitter’s crackdown on interference operations after the 2016 US Presidential election. By combining text embedding and transfer learning, our framework can detect, with promising accuracy, malicious users posting in Tagalog without any prior knowledge or training on malicious content in that language. We first learn an embedding model for each language, namely a high-resource language (English) and a low-resource one (Tagalog), independently. Then, we learn a mapping between the two latent spaces to transfer the detection model. We demonstrate that the proposed approach significantly outperforms state-of-the-art models and yields marked advantages in settings with very limited training data—the norm when dealing with detecting malicious activity in online platforms.
- 30 April 2023
- Published by Association for Computing Machinery (ACM)
- p. 1358-1364
Open AccessDetecting Social Media Manipulation in Low-Resource Languages
Social media have been deliberately used for malicious purposes, including political manipulation and disinformation. Most research focuses on high-resource languages. However, malicious actors share content across countries and languages, including low-resource ones. Here, we investigate whether and to what extent malicious actors can be detected in low-resource language settings. We discovered that a high number of accounts posting in Tagalog were suspended as part of Twitter’s crackdown on interference operations after the 2016 US Presidential election. By combining text embedding and transfer learning, our framework can detect, with promising accuracy, malicious users posting in Tagalog without any prior knowledge or training on malicious content in that language. We first learn an embedding model for each language, namely a high-resource language (English) and a low-resource one (Tagalog), independently. Then, we learn a mapping between the two latent spaces to transfer the detection model. We demonstrate that the proposed approach significantly outperforms state-of-the-art models and yields marked advantages in settings with very limited training data—the norm when dealing with detecting malicious activity in online platforms.
- Open Access
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 538-547
Investigating the Representation of Open Domain Dialogue Context for Transformer Models
Vishakh Padmakumar, Behnam Hedayatnia, Di Jin, Patrick Lange, Seokhwan Kim, Nanyun Peng, Yang Liu, Dilek Hakkani-Tur. Proceedings of the 24th Meeting of the Special Interest Group on Discourse and Dialogue. 2023.
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 538-547
Open AccessInvestigating the Representation of Open Domain Dialogue Context for Transformer Models
Vishakh Padmakumar, Behnam Hedayatnia, Di Jin, Patrick Lange, Seokhwan Kim, Nanyun Peng, Yang Liu, Dilek Hakkani-Tur. Proceedings of the 24th Meeting of the Special Interest Group on Discourse and Dialogue. 2023.
- Open Access
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 5137-5151
Code-Switched Text Synthesis in Unseen Language Pairs
I-Hung Hsu, Avik Ray, Shubham Garg, Nanyun Peng, Jing Huang. Findings of the Association for Computational Linguistics: ACL 2023. 2023.
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 5137-5151
Open AccessCode-Switched Text Synthesis in Unseen Language Pairs
I-Hung Hsu, Avik Ray, Shubham Garg, Nanyun Peng, Jing Huang. Findings of the Association for Computational Linguistics: ACL 2023. 2023.
- Top Cited PapersOpen Access
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 3378-3465
DOC: Improving Long Story Coherence With Detailed Outline Control
Kevin Yang, Dan Klein, Nanyun Peng, Yuandong Tian. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 3378-3465
Top Cited PapersOpen AccessDOC: Improving Long Story Coherence With Detailed Outline Control
Kevin Yang, Dan Klein, Nanyun Peng, Yuandong Tian. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.