Danyang Liu

About Me

I am a Ph.D. student at the UKRI CDT in NLP at University of Edinburgh, advised by Prof. Frank Keller and Prof. Mirella Lapata. I am also a member of the Edinburgh NLP of School of Informatics. I interned at Amazon Prime Video in 2024 and Huawei Edinburgh Research Center in 2021.

My research interests lie at the intersection of language and vision, specifically on:

• Controllable and visually-grounded text generation from visual input
• Reasoning and distillation in large vision-language models (VLMs)
• Understanding, generation, and evaluation of stories and narrative texts

Seeking postdoc or industry research roles in the UK for 2025. Please reach out with opportunities!

Preprints

• Detect, Disambiguate, and Translate: On-Demand Visual Reasoning for Multimodal Machine Translation with Large Vision-Language Models
  Danyang Liu, Fanjie Kong, Xiaohang Sun, Dhruva Patil, Zhu Liu, Vimal Bhat, Najmeh Sadoughi
  
• Generating Visual Stories with Grounded and Coreferent Characters
  Danyang Liu, Mirella Lapata, Frank Keller
  

Publications

• TrustScore: Reference-Free Evaluation of LLM Response Trustworthiness
  Danna Zheng, Danyang Liu, Mirella Lapata, Jeff Z. Pan
  Secure and Trustworthy Large Language Models Workshop (SeT-LLM @ ICLR), 2024.
  
• Visual Storytelling with Question-Answer Plans
  Danyang Liu, Mirella Lapata, Frank Keller
  Empirical Methods in Natural Language Processing (EMNLP Findings), 2023.
  
• Detecting and Grounding Important Characters in Visual Stories
  Danyang Liu and Frank Keller
  AAAI Conference on Artificial Intelligence (AAAI), 2023.
  
• Automatically Discarding Straplines to Improve Data Quality for Abstractive News Summarization
  Amr Keleg*, Matthias Lindemann*, Danyang Liu*, Wanqiu Long*, Bonnie L. Webber
  (*=Equal contribution)
  NLP Power! The First Workshop on Efficient Benchmarking (ACL Workshop), 2022.
  
• A Character-Centric Neural Model for Automated Story Generation
  Danyang Liu, Juntao Li, Meng-Hsuan Yu, Dongyan Zhao, Rui Yan
  AAAI Conference on Artificial Intelligence (AAAI), 2020.
  
• Draft and Edit: Automatic Storytelling Through Multi-Pass Hierarchical Conditional Variational Autoencoder
  Meng-Hsuan Yu, Juntao Li, Danyang Liu, Dongyan Zhao, Rui Yan
  AAAI Conference on Artificial Intelligence (AAAI), 2020.
  
• A Transformer-Based Variational Auto Encoder for Sentence Generation
  Danyang Liu and Gongshen Liu
  International Joint Conference on Neural Networks (IJCNN), 2019.

Background

• PhD, University of Edinburgh, Edinburgh, U.K. 2025 (expected)
  
• Master, Shanghai Jiao Tong University, Shanghai, China. 2020
  
• Bachelor, Southeast University, Nanjing, China. 2017

Last updated in Oct, 2024.