Research

Combinatorial Inference


Combinatorial inference develops novel theories on high dimensional universality phenomenon and combinatorial information-theoretical lower bounds and proposes methods to conduct the global inferential analysis on the topological properties of the graphs in the graphical models and other combinatorial structures.

Inference on the optimal assortment in the multinomial logit model
X. Chen S. Shen, E. Fang, J. Lu
ACM Conference on Economics and Computation, 2023.
Combinatorial-Probabilistic Trade-Off: Community Properties Test in the Stochastic Block Models
S. Shen, J. Lu
International Conference on Learning Representations (spotlight paper), 2023.
StarTrek: Combinatorial Variable Selection with False Discovery Rate Control
L. Zhang and J. Lu
Under revision, 2023.
Computational and Statistical Tradeoffs in Inferring Combinatorial Structures of Ising Model
J. Ying, Z. Wang, J. Lu
In International Conference on Machine Learning, pp. 4901-4910
Lagrangian Inference for Ranking Problems
Y. Liu, E.X. Fang, J. Lu
Operations Research, 2022.
ARCH: Large-scale Knowledge Graph via Aggregated Narrative Codified Health Records Analysis
Z. Gan, D. Zhou, E. Rush, V. Panickan, Y. Ho, G. Ostrouchov, Z. Xu, S. Shen, X. Xiong, K. Greco, C. Hong, C. Bonzel, J. Wen, L. Costa, T. Cai, E. Begoli, Z. Xia, M G, K. Liao, K. Cho, T. Cai, J. Lu
Submitted, 2023.
Knowledge Graph Embedding with Electronic Health Records Data via Latent Graphical Block Model
J. Lu, T. Cai
Submitted, 2023.
Combinatorial Inference for Graphical Models
Matey Neykov*, Junwei Lu*, Han Liu (*: equal contribution)
Annals of Statistics, to appear.
[Arxiv]
Adaptive Inferential Method for Monotone Graph Invariants
Junwei Lu, Matey Neykov, Han Liu
Submitted, 2016.
[Arxiv] [R package]
ICSA 2017 Student Paper Award
Post-Regularization Inference for Dynamic Nonparanormal Graphical Models
Junwei Lu, Mladen Kolar, Han Liu
Journal of Machine Learning Research, to appear.
[Arxiv]
Graphical Fermat's Principle and Triangle-Free Graph Estimation
Junwei Lu, Han Liu
Under revision at Journal of Machine Learning Research, 2016.
[Arxiv]

Complex Data Inference


Modern statistical methods deal with datasets with complex structures including high-dimensionality, heterogeneity, heavy-tailness, time-dependency and so on. Complex data inference aims to develop a new generation of inferential methods like hypothesis testing, confidence intervals and false discovery control for complex datasets.

High Dimensional Inference


Knowledge Graph Embedding with Electronic Health Records Data via Latent Graphical Block Model
J. Lu, T. Cai
Submitted, 2023.
Multi-source Learning via Completion of Block-wise Overlapping Noisy Matrices
D. Zhou, T. Cai, J. Lu
Under revision, 2023.
Graph over-parameterization: Why the graph helps the training of deep graph convolutional network
Y. Lin, S Li, J Xu, J Xu, D Huang, W Zheng, Y Cao, J. Lu
Neurocomputing, 2023.
Multimodal representation learning for predicting molecule–disease relations
J Wen, X Zhang, E Rush, V A Panickan, X Li, T Cai, D Zhou, Y Ho, L Costa, E Begoli, C Hong, J Gaziano, K Cho, J. Lu, K Liao, M Zitnik, T Cai
Bioinformatics, 2023.
Multiview Incomplete Knowledge Graph Integration with Application to Cross-institutional EHR Data Harmonization
D. Zhou, Z. Gan, X. Shi, A. Patwari, E. Rush, CL. Bonzel, V. A. Panickan, C. Hong, YL. Ho, T. Cai, L. Costa, X. Li, V.M. Castro, S.N. Murphy, G. Brat, G. Weber, P. Avillach, J.M. Gaziano, K. Cho, K. Liao, J. Lu*, T. Cai* (*: co-senior author)
Journal of Biomedical Informatics, 2022
Inter-Subject Analysis: Inferring Sparse Interactions with Dense Intra-Graphs
Cong Ma, Junwei Lu, Han Liu
Submitted, 2017.
[Arxiv]
ICSA 2017 Student Paper Award

Heterogenous Data


Nonparametric Heterogeneity Testing For Massive Data
Junwei Lu, Guang Cheng, Han Liu
Under revision at Bernoulli, 2017.
[Arxiv]

Robustness


Robust Scatter Matrix Estimation for High Dimensional Distributions with Heavy Tails
Junwei Lu, Fang Han, Han Liu
Submitted, 2016.
[PDF]

Time Series


Sparse Principal Component Analysis in Frequency Domain for Time Series
Junwei Lu, Yichen Chen, Xiuneng Zhu, Fang Han, Han Liu
Submitted, 2016.
[PDF]

Complicated Algorithms Inference


Modern data analysis introduces novel estimation methods including distributed algorithms, privacy methods, kernel estimators, etc. We aims to develop inference method to handle the uncertainty assessment under these algorithms.

Distributed Data Analysis


Distributed Testing and Estimation under Sparse High Dimensional Models
Heather Battey, Jianqing Fan, Han Liu, Junwei Lu, Ziwei Zhu (alphabetical order)
Annals of Statistics, to appear.
[Arxiv]

Sparse Additive Model


Kernel Meets Sieve: Post-Regularization Confidence Bands for Sparse Additive Model
Junwei Lu, Mladen Kolar, Han Liu
Submitted, 2016.
[Arxiv] [PDF]
ASA Best Student Paper in Nonparametric Statistics Finalist

Statistical Optimization


The nature of statistical problems brings new insights into our understanding of the geometry of optimization problems. Utilizing these insights, we are able to study the theorectial perfomance of nonconvex optimization problems and propose new algorithms.

Fast Distributed Principal Component Analysis of Large-Scale Federated Data
S. Shen, J. Lu, X. Lin
Submitted, 2023.
Federated Offline Reinforcement Learning
D. Zhou, Y. Zhang, A. Sonabend, Z. Wang, J. Lu, T. Cai
Under revision, 2023.
Expert-Supervised Reinforcement Learning for Offline Policy Learning and Evaluation
A. Sonabend W., J. Lu, L.A. Celi, T. Cai, Peter Szolovits
NeurIPS 2020.
Computational and Statistical Tradeoffs in Inferring Combinatorial Structures of Ising Model
J. Ying, Z. Wang, J. Lu
In International Conference on Machine Learning, pp. 4901-4910
Provable Sparse Tensor Decomposition
Wei Sun, Junwei Lu, Han Liu, Guang Cheng
Journal of the Royal Statistical Society: Series B, 2016.
[Arxiv][Link]
Symmetry, Saddle Points, and Global Geometry of Nonconvex Matrix Factorization
Xingguo Li, Zhaoran Wang, Junwei Lu, Raman Arora, Jarvis Haupt, Han Liu, Tuo Zhao
Under revision at IEEE Transactions on Information Theory, 2017.
[Arxiv]
Application of the Strictly Contractive Peaceman-Rachford Splitting Method to Multi-block Separable Convex Programming
Bingsheng He, Han Liu, Junwei Lu, Xiaoming Yuan (alphabetical order)
Splitting Methods in Communication, Imaging, Science, and Engineering (In Roland Glowinski, Stanley J. Osher, Wotao Yin (Eds.)), Springer, 2017.
[Optimization Online]