About

My name is Nan Jiang (姜楠). I’m a final year PhD student at the Department of Computer Science, at Purdue University. I am fortunate to be supervised by Professor Yexiang Xue. Previously, I obtained my master’s degree at Beihang University under the supervision of Professor Wenge Rong and a bachelor’s degree at the Zhejiang University of Technology.

I’m currently actively looking for Postdoc positions. Please contact me if there are any available positions.

Research Statement and Teaching Statement

Contacts: Email / CV / Google Scholar / LinkedIn

Research Interests

My PhD research focuses on Machine Learning with Constraint Reasoning and symbolic regression for scientific discovery.

Research Highlights

AI-driven Scientific Discovery

  • We propose an end-to-end framework to learn physics models in the form of Partial Differential Equations (PDEs) directly from the experiment data.
  • We scale up learning first-principle models harnessing randomized algorithms, exploiting the fact that the temporal evolutions of many physical systems often consist of gradually changing updates across wide areas in addition to a few rapid updates concentrated in a small set of “interfacial” regions.
  • The development of AI-driven scientific discovery approaches was motivated by the real-world application of learning the physics model of nano-scale crystalline defects in materials under extreme conditions.
  • Papers:

Constraint Reasoning Embedded in Machine Learning

  • We propose COnstraint REasoning embedded structured learning (CORE), a scalable constraint reasoning and machine learning integrated approach for learning over structured domains.
  • We embed decision diagrams, a popular constraint reasoning tool, as a fully-differentiable module into deep learning models.
  • In data-driven operational research and program synthesis from the natural language, the structures generated with CORE satisfy 100% of the constraints when using exact decision diagrams. In addition, CORE boosts learning performance by reducing the modeling space via constraint satisfaction. CORE also generates designs that satisfy complex user specifications as well as meet aesthetics, utility, and convenience requirements.

Publications

  1. Solving Satisfiability Modulo Counting for Symbolic and Statistical AI Integration with Provable Guarantees.
    Jinzhao Li, Nan Jiang and Yexiang Xue.
    AAAI 2024. (pdf)(poster)[https://github.com/jil016/xor-smc]

  2. Racing Control Variable Genetic Programming for Symbolic Regression.
    Nan Jiang, Yexiang Xue.
    AAAI 2024. (pdf)(poster)(code)[https://bitbucket.org/xlnxyx/racing_cvgp/src/master/]

  3. Symbolic Regression via Control Variable Genetic Programming.
    Nan Jiang, Yexiang Xue.
    ECML-PKDD 2023. (pdf) (poster) (slides) (code)

  4. Learning Markov Random Fields for Combinatorial Structures via Sampling through Lovász Local Lemma.
    Nan Jiang, Yi Gu, Yexiang Xue. (First two authors contributed equally)
    AAAI 2023. (pdf) (poster) (code)

  5. Constraint Reasoning Embedded in Structural Prediction.
    Nan Jiang, Maosen Zhang, Willem-Jan van Hoeve, Yexiang Xue.
    JMLR 2022. (pdf) (poster) (code)

  6. Massive Text Normalization via an Efficient Randomized Algorithm.
    Nan Jiang, Chen Luo, Vihan Lakshman, Yesh Dattatreya, Yexiang Xue.
    TheWebConf 2022. (pdf) (poster)[](code)

  7. PALM: Probabilistic Area Loss Minimization for Protein Sequence Alignment.
    Fan Ding, Nan Jiang, Jianzhu Ma, Jian Peng, Jinbo Xu, and Yexiang Xue. (First two authors contributed equally)
    UAI 2021. (pdf) (poster) (code)

  8. Language Generation via Combinatorial Constraint Satisfaction: A Tree Search Enhanced Monte-Carlo Approach.
    Maosen Zhang, Nan Jiang, Lei Li, and Yexiang Xue.
    Finding in EMNLP 2020. (pdf) (poster) (code)

  9. A dual channel class hierarchy based recurrent language modeling.
    Libin Shi, Wenge Rong, Shijie Zhou, Nan Jiang, Zhang Xiong.
    Neurocomputing 2020.

  10. LSDSCC: A Large Scale Domain-Specific Conversational Corpusfor Response Generation with Diversity Oriented Evaluation Metrics.
    Zhen Xu, Nan Jiang and et. al.
    NAACL-HLT 2018. (pdf)

  11. Exploration of Tree-based Hierarchical Softmax for Recurrent Language Models.
    Nan Jiang, Wenge Rong, Min Gao, Yikang Shen, Zhang Xiong.
    IJCAI 2017. (pdf) (code)

  12. Event Trigger Identification with Noise Contrastive Estimation.
    Nan Jiang, Wenge Rong, Yifan Nie, Yikang Shen, and Zhang Xiong.
    IEEE/ACM Transactions on Computational Biology and Bioinformatics, 2017. (pdf) (code)

  13. An Empirical Analysis of Different Sparse Penalties for Autoencoder in Unsupervised Feature Learning.
    Nan Jiang, Wenge Rong, Baolin Peng, Yifan Nie, Zhang Xiong.
    IJCNN, 2015. (pdf) (code)[https://github.com/jiangnanhugo/Undergraduate_Design/tree/master/Self-Taught-Learning]