Dou Hu

Dou Hu

I am currently a Ph.D. candidate at the University of Chinese Academy of Sciences (UCAS) and the Institute of Information Engineering, Chinese Academy of Sciences (IIE, CAS), under the supervision of Prof. Songlin Hu. I completed my bachelor’s degree in Physics from Shenzhen University and my master’s degree in Software Engineering from National Computer System Engineering Research Institute of China (a joint program with UCAS). Before starting my Ph.D., I also worked as an algorithm researcher at Ping An.

My research interests include natural language processing and machine learning. And my research goal is to enable AI to better understand and generate human language. To achieve this, I am currently focused on developing principled representation learning techniques to enhance the generalization, robustness, efficiency, and trustworthiness of language models.

Research Interests

  • Representation Learning
    • Learning theory & principle, generalization and robustness, multi-task representation learning
  • Sentiment Analysis
    • Sentiment classification, emotion recognition, dimensional emotion analysis, irony detection, stance detection
  • Trustworthy Machine Learning
    • Trustworthy large language models (truthfulness, safety, fairness), misinformation detection, harmful content detection

🔥 News

  • Dec. 2024: 🎉 One paper is accepted at AAAI 2025.
  • Dec. 2024: 🎖 I received the Director Scholarship (Excellent Prize).
  • Nov. 2024: 🎖 I received the National Scholarship.
  • May. 2024: 🎉 One paper is accepted at ACL 2024.
  • Feb. 2024: 🎉 One paper is accepted at LREC-COLING 2024. Congrats to Yinan Bao!
  • Dec. 2023: 🎉 One paper is accepted at ICASSP 2024. Congrats to Lingwei Wei!
  • Dec. 2023: 🎉 One paper is accepted at AAAI 2024 (Oral).
  • Dec. 2023: 📚 I served as an Area Chair for ACL ARR 2023.
  • Jul. 2023: 🏆 I received the Best System Award at SemEval 2023 in Task 12.
  • May. 2023: 🎉🎉 Two papers are accepted at ACL 2023.

📖 Selected Publications (Full List)

First-author Publications:

  • An Information-theoretic Multi-task Representation Learning Framework for Natural Language Understanding.
    Dou Hu, Lingwei Wei, Wei Zhou, Songlin Hu. AAAI 2025. (CCF-A)
    [Code]

  • Representation Learning with Conditional Information Flow Maximization.
    Dou Hu, Lingwei Wei, Wei Zhou, Songlin Hu. ACL 2024. (CCF-A)
    [Paper] [Code]

  • Structured Probabilistic Coding.
    Dou Hu, Lingwei Wei, Yaxin Liu, Wei Zhou, Songlin Hu. AAAI 2024. (CCF-A, Oral 2.2%)
    [Paper] [Code]

  • Supervised Adversarial Contrastive Learning for Emotion Recognition in Conversations.
    Dou Hu, Yinan Bao, Lingwei Wei, Wei Zhou, Songlin Hu. ACL 2023. (CCF-A)
    [Paper] [Code]

  • UCAS-IIE-NLP at SemEval-2023 Task 12: Enhancing Generalization of Multilingual BERT for Low-resource Sentiment Analysis.
    Dou Hu, Lingwei Wei, Yaxin Liu, Wei Zhou, Songlin Hu. SemEval@ACL 2023. (First Prize & Best System Award)
    [Paper] [Code]

  • VarMAE: Pre-training of Variational Masked Autoencoder for Domain-adaptive Language Understanding.
    Dou Hu, Xiaolong Hou, Xiyang Du, Mengyuan Zhou, Lianxin Jiang, Yang Mo, and Xiaofeng Shi. Findings of EMNLP 2022. (CCF-B Findings)
    [Paper]

  • MM-DFN: Multimodal Dynamic Fusion Network for Emotion Recognition in Conversations.
    Dou Hu, Xiaolong Hou, Lingwei Wei, Lianxin Jiang, Yang Mo. ICASSP 2022. (CCF-B)
    [Paper] [Code]

  • PALI-NLP at SemEval-2022 Task 4: Discriminative Fine-tuning of Transformers for Patronizing and Condescending Language Detection.
    Dou Hu, Zhou Mengyuan, Xiyang Du, Mengfei Yuan, Jin Zhi, Lianxin Jiang, Mo Yang, Xiaofeng Shi. SemEval@NAACL 2022. (First Prize)
    [Paper]

  • DialogueCRN: Contextual Reasoning Networks for Emotion Recognition in Conversations.
    Dou Hu, Lingwei Wei, Xiaoyong Huai. ACL-IJCNLP 2021. (CCF-A, Oral)
    [Paper] [Code]

  • A Rumor Detection Approach based on Multi-relational Propagation Tree (一种基于多关系传播树的谣言检测方法).
    Dou Hu, Lingwei Wei, Wei Zhou, Xiaoyong Huai, Jizhong Han, Songlin Hu. Journal of Computer Research and Development 2021 (CCF-A Chinese)
    [Paper]

  • Hierarchical Interaction Networks with Rethinking Mechanism for Document-level Sentiment Analysis.
    Lingwei Wei*, Dou Hu*, Wei Zhou, Xuehai Tang, Xiaodan Zhang, Xin Wang, Jizhong Han, Songlin Hu. ECML-PKDD 2020. (CCF-B)
    [Paper] [Code] (* denotes co-first authors)

Co-author Publications:

  • Transferring Structure Knowledge: A New Task to Fake News Detection Towards Cold-Start Propagation.
    Lingwei Wei, Dou Hu, Wei Zhou, Songlin Hu. ICASSP 2024. (CCF-B)
    [Paper]

  • Multi-stream Information Fusion Framework for Emotional Support Conversation.
    Yinan Bao, Dou Hu, Lingwei Wei, Shuchong Wei, Wei Zhou, Songlin Hu. LREC-COLING 2024. (CCF-B)
    [Paper]

  • Modeling the Uncertainty of Information Propagation for Rumor Detection: A Neuro-Fuzzy Approach.
    Lingwei Wei, Dou Hu, Wei Zhou, Xin Wang, Songlin Hu. TNNLS 2024. (CCF-B, SCI-Q1)
    [Paper]

  • Modeling Both Intra-and Inter-Modality Uncertainty for Multimodal Fake News Detection.
    Lingwei Wei, Dou Hu, Wei Zhou, Songlin Hu. TMM 2023. (CCF-B, SCI-Q1)
    [Paper]

  • PALI-NLP at SemEval-2022 Task 6: iSarcasmEval- Fine-tuning the Pre-trained Model for Detecting Intended Sarcasm.
    Xiyang Du, Dou Hu, Meizhi Jin, Lianxin Jiang, Xiaofeng Shi. SemEval@NAACL 2022. (First Prize)
    [Paper]

  • PALI at SemEval-2022 Task 7: Identifying Plausible Clarifications of Implicit and Underspecified Phrases in Instructional Texts.
    Mengyuan Zhou, Dou Hu, Mengfei Yuan, Meizhi Jin, Xiyang Du, Lianxin Jiang, Yang Mo, Xiaofeng Shi. SemEval@NAACL 2022. (First Prize)
    [Paper]

  • Towards Propagation Uncertainty: Edge-enhanced Bayesian Graph Convolutional Networks for Rumor Detection.
    Lingwei Wei, Dou Hu, Wei Zhou, Zhaojuan Yue, Songlin Hu. ACL-IJCNLP 2021. (CCF-A, Oral)
    [Paper]

📈 Industry Applications

  • Galaxy Generative AI Safety Evaluation Platform, IIE, CAS, 2023-2025.
    We developed a comprehensive safety evaluation platform for evaluating LLMs in China. The platform leverages three advanced representation learning techniques, including Supervised Adversarial Contrastive Learning (Hu et al., 2023a), Structured Probabilistic Coding (Hu et al., 2023b), and Conditional Information Flow Maximization (Hu et al., 2024), to achieve highly generalized and effective risk identification in generated content. It has generated over 100 automated risk reports for more than 60 mainstream LLMs in China, with detection performance significantly surpassing that of Google/OpenAI APIs.

  • AI Cloud Interview Platform, Ping An Life Insurance, 2021-2022.
    We developed an AI interview platform for training recruitment and employee. The platform use multi-task learning techniques for efficient model training and inference, enabling better intent understanding and interview evaluation. It has supported over 20 million interviews, totaling 3.4 million hours, and provided annual support for 4.72 million recruitment interviews.

  • Intelligent Visit Assistant Tool, Ping An Life Insurance, 2021-2022.
    We developed an AI-assisted visit tool for insurance sales. The tool leveraged the domain-specific pre-training technique Variational Masked Autoencoder (Hu et al., 2022) to enhance the language understanding capabilities of general PLMs for financial domain data. It has provided intelligent assistance and visit summaries for over 1 million agents during client interactions, achieving more than 10 million online client contacts annually.

🎖 Honors and Awards

  • Baidu Scholarship (Global Top 40), Baidu, 2024
  • Director Scholarship (Excellent Prize), IIE CAS, 2024
  • National Scholarship, China’s Ministry of Education, 2024
  • Pacemaker to Merit Student, University of Chinese Academy of Sciences, 2023-2024
  • Merit Student, University of Chinese Academy of Sciences, 2023-2024
  • Best System Award, Afrisenti-SemEval 2023 (Primary Contributor)
  • First Prize, Sentiment Analysis for African Languages Task (1/213, Team Leader), SemEval 2023
  • First Prize, Patronizing and Condescending Language Detection Task (1/300, Team Leader), SemEval 2022
  • First Prize, Intended Sarcasm Detection Task, SemEval 2022
  • First Prize, Multimedia Automatic Misogyny Identification Task, SemEval 2022
  • Runner-up, Identifying Plausible Clarifications Task, SemEval 2022

💬 Invited Talks

  • Dec. 2024, Invited talk, “Information-theoretic Representation learning for Natural Language Understanding” (面向自然语言理解的信息论表示学习), School of Artificial Intelligence, Beijing Normal University.
  • Jan. 2024, Pre-talk Presentation, “Structured Probabilistic Coding”, Youth Working Committee, Chinese Information Processing Society of China (CIPS).
  • Jan. 2024, Youth PhD Talk, “Structured Probabilistic Coding”, AI TIME.
  • Sep. 2023, Youth PhD Talk, “Supervised Adversarial Contrastive Learning”, AI TIME.

📝 Academic Services

  • Area Chair: ACL ARR 2023
  • Conference Program Committee / Reviewer: ACL 2023/2024, EMNLP 2023/2024, NAACL 2024, NeurIPS 2024, KDD 2025, ICLR 2025, ICML 2025, etc.
  • Journal Reviewer: IEEE Transactions on Multimedia (TMM), IEEE Transactions on Circuits and Systems for Video Technology (TCSVT), Engineering Applications of Artificial Intelligence (EAAI)