About Me ([GitHub] [Google Scholar])

I am currently a final-year PhD candidate at School of Artificial Intelligence, Nanjing University (NJU), and a member of NJU-NLP Research Group. My recent works are mainly on in-context learning, multi-modal information extraction, multi-modal large language models, and NLP applications.

Latest News

  • 2025-02-22: One paper on Efficient Multimodal Large Language Models is accepted by ICLR 2025.
  • 2024-09-20: One paper on Multimodal Hallucination Mitigation is accepted by EMNLP 2024.
  • 2024-05-24: Release a paper on Multi-modal Large Language Models to Arxiv.
  • 2024-05-16: Two papers are accepted by Findings of ACL 2024.
  • 2024-02-16: Release a paper on Multimodal Hallucination Mitigation to Arxiv.
  • 2023-10-08: One paper on multimodal ABSA is accepted by EMNLP 2023.
  • 2023-10-03: Release a paper on in-context learning to Arxiv.
  • 2023-07-26: One paper on multimodal entity linking is accepted by ACM Multimedia 2023.
  • 2023-05-11: One paper is accepted by TASLP 2023.
  • 2023-05-02: One paper is accepted by Findings of ACL 2023.
  • 2022-10-06: One paper on few-shot aspect category detection is accepted by Findings of EMNLP 2022.
  • 2022-08-17: One paper on multimodal ABSA is accepted by COLING 2022.
  • 2022-06-30: One paper on multimodal NER is accepted by ACM Multimedia 2022.
  • 2020-10-01: One paper on ABSA is accepted by COLING 2020.
  • 2020-09-14: One paper on aspect triplet extraction is accepted by Findings of EMNLP 2020.
  • 2019-11-11: One paper on TOWE is accepted by AAAI 2020.

Pre-prints

AlignGPT: Multi-modal Large Language Models with Adaptive Alignment Capability
Fei Zhao*, Taotian Pang*, Chunhui Li, Zhen Wu, Junjie Guo, Shangyu Xing, and Xinyu Dai
[Project page] [Paper] [Code] [Model] [Demo]
AlignGPT generates different levels of alignment capabilities in the pre-training stage, and then adaptively combines these alignment capabilities in the instruction-tuning stage to meet the alignment needs of different instructions.
Dynamic Demonstrations Controller for In-Context Learning
Fei Zhao, Taotian Pang, Zhen Wu, Zheng Ma, Shujian Huang, Xinyu Dai
Under review
[Paper] [Code] [中文解读] [BibTex]
We propose a method named D2Controller, which not only boosts ICL performance but also saves time and space during inference of the LLMs.

Selected Publications ([Full List])

(* indicates equal contribution)
EFUF: Efficient Fine-grained Unlearning Framework for Mitigating Hallucinations in Multimodal Large Language Models
Shangyu Xing, Fei Zhao, Zhen Wu, Tuo An, Weihao Chen, Chunhui Li, Jianbing Zhang, Xinyu Dai
EMNLP, 2024
[Paper] [Code] [中文解读] [BibTex]
We propose an efficient fine-grained unlearning framework EFUF, which can obtain positive and negative examples separately in a cost-effective manner.
M2DF: Multi-grained Multi-curriculum Denoising Framework for Multimodal Aspect-based Sentiment Analysis
Fei Zhao*, Chunhui Li*, Zhen Wu, Yawen Ouyang, Jianbing Zhang and Xinyu Dai
EMNLP, 2023
[Paper] [Code] [中文解读] [BibTex]
In this work, we focus on whether the negative impact of noisy images can be reduced without filtering the data. To achieve this goal, we propose a Multi-grained Multi-curriculum Denoising Framework.
DRIN: Dynamic Relation Interactive Network for Multimodal Entity Linking
Shangyu Xing*, Fei Zhao*, Zhen Wu, Chunhui Li, Jianbing Zhang and Xinyu Dai
ACM MM, 2023
[Paper] [Code] [中文解读] [BibTex]
DRIN explicitly models four different types of alignment between a mention and entity and builds a dynamic Graph Convolutional Network (GCN) to select the different alignment relations for different input samples.
Label-Driven Denoising Framework for Multi-Label Few-Shot Aspect Category Detection
Fei Zhao*, Yuchen Shen*, Zhen Wu, Xinyu Dai
EMNLP findings, 2022
[Paper] [Slides] [Code] [中文解读] [BibTex]
We propose a novel Label-Driven Denoising Framework (LDF) to alleviate the noise problems for the FS-ACD task.
Learning from Different text-image Pairs: A Relation-enhanced Graph Convolutional Network for Multimodal NER
Fei Zhao, Chunhui Li, Zhen Wu, Shangyu Xing, Xinyu Dai
ACM MM, 2022
[Paper] [Slides] [Code] [中文解读] [BibTex]
We propose leveraging the external matching relations between different (text, image) pairs to improve the performance on the MNER task.
Latent Opinions Transfer Network for Target-Oriented Opinion Words Extraction
Zhen Wu*, Fei Zhao*, Xin-Yu Dai, Shujian Huang, Jiajun Chen
AAAI, 2020 (oral presentation)
[Paper] [Slides] [Code] [中文解读] [BibTex]
We transter latent opinion knowledge from resource-rich review sentiment classification datasets to low-resource TOWE task.

Journal Papers

Label-correction Capsule Network for Hierarchical Text Classification
Fei Zhao, Zhen Wu, Liang He, Xinyu Dai
IEEE Transactions on Audio, Speech and Language Processing(TASLP), 2023
[Paper] [Slides] [Code] [中文解读] [BibTex]
We design two novel approaches to weaken the impact of incorrect parent-level labels on the child-level classification.

Experience

  • Jun. 2024 - Nov. 2024, research intern at Xiaohongshu.
  • Apr. 2020 - Nov. 2020, research intern at Tencent WeChat Group.

Honor

  • National Scholarship, 2018.
  • Outstanding Graduate, 2018.
  • Outstanding Undergraduation Thesis, 2018.
  • Huawei Scholarship, 2020.
  • Huawei Scholarship, 2022.
  • First-Class Yingcai Scholarship, 2023.

Teaching Assistant

  • Natural Language Processing (For graduate students, Fall, 2021)
  • Natural Language Processing (For undergraduate students, Spring, 2022)

Review Services

Program Committee Member/Conference Reviewer
International Conference on Computer Vision (ICCV), 2025
Forty-Second International Conference on Machine Learning (ICML), 2025
International World Wide Web Conference (WWW), 2025
The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025
International Conference on Learning Representations (ICLR), 2025
Neural Information Processing Systems (NIPS), 2024
Conference on Language Modeling (COLM), 2024
ACL Rolling Review 2021, 2022, 2023, 2024
ACM Multimedia (ACM MM), 2023
Empirical Methods in Natural Language Processing (EMNLP), 2022, 2023
AAAI Conference on Artificial Intelligence (AAAI), 2021, 2022, 2023, 2024
Annual Meeting of the Association for Computational Linguistics (ACL), 2021, 2023
North American Chapter of the Association for Computational Linguistics (NAACL), 2022
International Workshop on Natural Language Processing for Social Media (SocialNLP), 2021