Welcome to Dixin Luo's Homepage

Dixin Luo （罗迪新）

Assistant Professor

School of Computer Science and Technology, Beijing Institute of Technology

dxluo611 (at) gmail (dot) com

Biography

My research is situated in data mining and machine learning, especially graph analysis, machine learning for healthcare, point process-based model, time series analysis and Markovian model.

I received my Ph. D. degree in Electronic Engineering from Shanghai Jiao Tong University (SJTU), Shanghai, China, in 2016. I was jointly supervised by Prof. Wenjun Zhang (SJTU), Prof. Hongyuan Zha (Georgia Tech) and Prof. Xiaokang Yang (SJTU). Before that, I received my B.E. degree in Electronic Engineering from Shanghai Jiao Tong University in 2010.

Work Experience

  • Assistant Professor Since Nov 2020

Computer Science, Beijing Institute of Technology

  • Postdoctoral Associate Feb 2018 to Nov 2020

ECE, Duke University

Supervisor: Prof. Lawrence Carin

  • Postdoctoral Fellow Jul 2016 to Jun 2017

Faculty of Information, University of Toronto

Supervisor: Prof. Kelly Lyons

  • Teaching Assistant Sep 2016 – Dec 2016

Faculty of Information, University of Toronto

Course Name: Data Analytics: Introduction, Methods and Practical Approaches

Instructor: Prof. Periklis Andritsos

  • Co-chair November 02, 2016

The Workshop on Data-driven Knowledge Mobilization, CASCON 2016

Chairs: Kelly Lyons, Eleni Stroulia, Dixin Luo, Renee Miller, Vio Onut

  • Summer Intern Jul 2015 – Sep 2015

IBM China Research Lab, Shanghai

Recommendation Systems

Mentor: Junchi Yan

Publications

  • Hongteng Xu, Dixin Luo, Lawrence Carin, Hongyuan Zha: Learning Graphons via Structured Gromov-Wasserstein Barycenters. AAAI 2021.

  • Dixin Luo, Hongteng Xu, Lawrence Carin: Hierarchical Optimal Transport for Robust Multi-ViewLearning. arXiv preprint arXiv:2006.03160. 2020

  • Dixin Luo, Hongteng Xu, Ricardo Henao, Svati Shah, Lawrence Carin: Learning Autoencoders with Relational Regularization. ICML 2020.

  • Dixin Luo, Hongteng Xu, Lawrence Carin: Fused Gromov-Wasserstein Alignment for Hawkes Processes. Learning with Temporal Point Processes, NeurIPS 2019 Workshop.

  • Hongteng Xu, Dixin Luo, Lawrence Carin: Scalable Gromov-Wasserstein Learning for Graph Partitioning and Matching. NeurIPS 2019.

  • Dixin Luo, Hongteng Xu, Lawrence Carin: Interpretable ICD Code Embeddings with Self- and Mutual-Attention Mechanisms. arXiv:1906.05492. 2019.

  • Dixin Luo, Hongteng Xu and Lawrence Carin: Adversarial Self-Paced Learning for Mixture Models of Hawkes Processes. ICML Time Series Workshop 2019.

  • Hongteng Xu, Dixin Luo, Hongyuan Zha, Lawrence Carin: Gromov-Wasserstein Learning for Graph Matching and Node Embedding. ICML 2019.

  • Hongteng Xu, Dixin Luo, Lawrence Carin: Online Continuous-Time Tensor Factorization Based on Pairwise Interactive Point Processes. IJCAI-ECAI 2018.

  • Hongteng Xu, Dixin Luo, Xu Chen and Lawrence Carin: Benefits from Superposed Hawkes Processes. AISTATS 2018.

  • Hongteng Xu, Dixin Luo, Hongyuan Zha: Learning Hawkes Processes from Short Doubly-Censored Event Sequences. ICML 2017.

  • Dixin Luo, Kelly Lyons: CASCONet: A Conference dataset. arXiv:1706.09485. 2017.

  • Dixin Luo, Hongteng Xu, Yi Zhen, Bistra N. Dilkina, Hongyuan Zha, Xiaokang Yang, Wenjun Zhang: Learning Mixtures of Markov Chains from Aggregate Data with Structural Constraints. IEEE Trans. Knowl. Data Eng. 28(6): 1518-1531 (2016)

  • Kelly A. Lyons, Eleni Stroulia, Dixin Luo, Renée J. Miller, Vio Onut: Data-driven knowledge mobilization. CASCON 2016: 280-282

  • Hongteng Xu, Licheng Yu, Dixin Luo, Hongyuan Zha, Yi Xu: Dictionary Learning with Mutually Reinforcing Group-Graph Structures. AAAI 2015: 3101-3107

  • Dixin Luo, Hongteng Xu, Yi Zhen, Xia Ning, Hongyuan Zha, Xiaokang Yang, Wenjun Zhang: Multi-Task Multi-Dimensional Hawkes Processes for Modeling Event Sequences. IJCAI 2015: 3685-3691

  • Dixin Luo, Hongteng Xu, Hongyuan Zha, Jun Du, Rong Xie, Xiaokang Yang, Wenjun Zhang: You Are What You Watch and When You Watch: Inferring Household Structures From IPTV Viewing Data. IEEE Transactions on Broadcasting 60(1): 61-72 (2014)

  • Hongteng Xu, Dixin Luo, Xiaoming Huo, Xiaokang Yang: World Expo Problem and Its Mixed Integer Programming Based Solution. BSI@PAKDD/BSIC@IJCAI 2013: 56-67

Project experience

  • Predictive Modeling for Biomedical Time-to-Event Data

July 2019-now Supervised by Prof. Ricardo Henao, Prof. Lawrence Carin

Working on the development of a predictive model for healthcare data.

1. Proposed an attention-based model for predicting patients will survive or not based on their healthcare data.

2. Proposed a multi-view learning-based method for predicting patients will survive or not robustly when some views are missing.

  • Multi-Source Activity Graph Latent Uncovering and Merging

Feb 2018 – now Supervised by Prof. Lawrence Carin

Working on the development of a time-evolving generative deep learning graph alignment algorithm.

1. Proposed a graph matching model based on Node2Vec and Coherent Point Drift (CPD) method.

2. Proposed a novel Gromov-Wasserstein learning framework to jointly match (align) graphs and learn embedding vectors for the associated graph nodes.

3. Proposed a scalable Gromov-Wasserstein learning (S-GWL) method and established a novel and theoretically-supported paradigm for large-scale graph analysis, especially for multi-graph partitioning and matching, based on Gromov-Wasserstein barycenter and optimal transport.

  • Machine Learning for Healthcare

Jan 2019 - Now Supervised by Prof. Lawrence Carin, Prof. Ricardo Henao

Proposed a novel and interpretable embedding method to represent the international statistical classification codes of diseases and related health problems (i.e., ICD codes), based on attention mechanism and optimal transport, and applied the method to procedure recommendation.

  • Data-Driven Knowledge Mobilization, Translation and Innovation

Jul 2016 – Jun 2017 Supervised by Prof. Kelly Lyons

1. Worked on predicting researchers’ future fundings according to their historical behaviors via a point process-based model;

2. Worked on co-authorship recommendation based on researchers’ historical publications and co-authorships.

3. Collected and constructed a new data set about academia conferences and analyzed their long-term dynamics and patterns;

(This data set, named ''CASCONet'', is shared on Github. https://github.com/iDBKMTI/CASCONet )

4. Built the project website via Wordpress: http://hypatia.cs.ualberta.ca/kmti/

  • IPTV User Behavior Analysis and System Simulation

Jun 2012 – Jan 2016

Proposed a Multi-task Multi-dimensional Hawkes Process (MMHP) for modeling event sequences where there exist multiple triggering patterns within sequences and structures across sequences.

  • E-commerce Recommendation

Jul 2015 – Sep 2015

Proposed a e-commerce recommendation model using a multi-dimensional Hawkes process model and nonnegative matrix factorization.

  • Crowd Analysis of Shanghai Expo 2010

Nov 2013 – Dec 2014

Proposed a method for learning mixtures of Markov chains (MMCs) from aggregate data with structural constraints.

Professional activities

  • Journal reviewer

IEEE Transactions on Knowledge and Data Engineering (TKDE), IEEE Transactions on Neural Networks and Learning Systems (TNNLS)

  • Conference reviewer

ICML 2020, ICML 2021, NeurIPS 2020, AAAI 2021

Skills

Python, Matlab, MySQL, Latex

Awards

University Outstanding Scholarship, Shanghai Jiao Tong University 2006-2008

University Honor Roll Student, Shanghai Jiao Tong University 2010-2012

Outstanding Paper Award of the Fourteenth Annual Meeting of Shanghai Institute of Communications 2013

Visiting Abroad Special Scholarship, Shanghai Jiao Tong University 2013-2014

Guanghua Scholarship, Shanghai Jiao Tong University 2014-2015