Hanjun Dai

Google Research

About Me

Sep 2019 - NOW


Aug 2014 - Aug 2019


Aug 2018 - Dec 2018


May 2018 - Aug 2018


May 2017 - Aug 2017


July 2013 - Apr 2014

Contact

hadai@google.com

Biography

I'm a senior research scientist at Google Brain.

I obtained my Ph.D. degree in CSE in Georgia Institute of Technology. My advisor is Prof. Le Song.

I received my B.S. in Computer Science, Fudan University in 2014. My advisor is Prof. Junping Zhang.

[ Google Scholar ] [ Github ]

Publications

Proceeding

  • Score-based Continuous-time Discrete Diffusion Models
    Haoran Sun, Lijun Yu, Bo Dai, Dale Schuurmans, Hanjun Dai
    International Conference on Learning Representations (ICLR) 2023.
    [ Paper ]

  • Any-scale Balanced Samplers for Discrete Space
    Haoran Sun, Bo Dai, Charles Sutton, Dale Schuurmans, Hanjun Dai
    International Conference on Learning Representations (ICLR) 2023.
    [ Paper ]

  • Learning to Optimize with Stochastic Dominance Constraints
    Hanjun Dai, Yuan Xue, Niao He, Bethany Wang, Na Li, Dale Schuurmans, Bo Dai
    International Conference on Artificial Intelligence and Statistics (AISTATS) 2023.
    [ arxiv ]

  • Discrete Langevin Samplers via Wasserstein Gradient Flow
    Haoran Sun, Hanjun Dai, Bo Dai, Haomin Zhou, Dale Schuurmans
    International Conference on Artificial Intelligence and Statistics (AISTATS) 2023.
    [ Paper ]

  • Optimal Scaling for Locally Balanced Proposals in Discrete Spaces
    Haoran Sun, Hanjun Dai, Dale Schuurmans
    Conference on Neural Information Processing Systems (NeurIPS) 2022.
    [ Paper ]

  • Does GNN Pretraining Help Molecular Representation?
    Ruoxi Sun, Hanjun Dai, Adams Wei Yu.
    Conference on Neural Information Processing Systems (NeurIPS) 2022.
    [ Paper ]

  • SMORE: Knowledge Graph Completion and Multi-hop Reasoning in Massive Knowledge Graphs.
    Hongyu Ren*, Hanjun Dai*, Bo Dai, Xinyun Chen, Denny Zhou, Jure Leskovec, Dale Schuurmans.
    ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD) 2022.
    [ arxiv ] [ Code ]

  • Marginal Distribution Adaptation for Discrete Sets via Module-Oriented Divergence Minimization
    Hanjun Dai, Mengjiao Yang, Yuan Xue, Dale Schuurmans, Bo Dai
    International Conference on Machine Learning (ICML) 2022.
    [ Paper ]

  • Path Auxiliary Proposal for MCMC in Discrete Space
    Haoran Sun, Hanjun Dai, Wei Xia, Arun Ramamurthy
    International Conference on Learning Representations (ICLR) 2022. Spotlight
    [ Paper ] [ Code ]

  • CodeTrek: Flexible Modeling of Code using an Extensible Relational Representation
    Pardis Pashakhanloo, Aaditya Naik, Yuepeng Wang, Hanjun Dai, Petros Maniatis, Mayur Naik
    International Conference on Learning Representations (ICLR) 2022.
    [ Paper ]

  • CrossBeam: Learning to Search in Bottom-Up Program Synthesis
    Kensen Shi*, Hanjun Dai*, Kevin Ellis, Charles Sutton
    International Conference on Learning Representations (ICLR) 2022.
    [ Paper ]

  • Neural Stochastic Dual Dynamic Programming
    Hanjun Dai*, Yuan Xue*, Zia Syed, Dale Schuurmans, Bo Dai
    International Conference on Learning Representations (ICLR) 2022.
    [ Paper ] [ arxiv ]

  • Combiner: Full Attention Transformer with Sparse Computation Cost
    Hongyu Ren*, Hanjun Dai*, Zihang Dai*, Mengjiao Yang, Jure Leskovec, Dale Schuurmans, Bo Dai
    Advances in Neural Information Processing Systems (NeurIPS) 2021. Spotlight
    [ arxiv ] [ Code ]

  • Towards understanding retrosynthesis by energy-based models
    Ruoxi Sun, Hanjun Dai, Li Li, Steven Kearnes, Bo Dai
    Advances in Neural Information Processing Systems (NeurIPS) 2021.
    [ Paper ]

  • SpreadsheetCoder: Formula Prediction from Semi-structured Context
    Xinyun Chen, Petros Maniatis, Rishabh Singh, Charles Sutton, Hanjun Dai, Max Lin, Denny Zhou
    International Conference on Machine Learning (ICML) 2021.
    [ arxiv ] [ Google AI Blog ]

  • LEGO: Latent Execution-Guided Reasoning for Multi-Hop Question Answering on Knowledge Graphs
    Hongyu Ren, Hanjun Dai, Bo Dai, Xinyun Chen, Michihiro Yasunaga, Haitian Sun, Dale Schuurmans,
    Jure Leskovec, Denny Zhou
    International Conference on Machine Learning (ICML) 2021.
    [ Paper ] [ Code ]

  • Molecule Optimization by Explainable Evolution
    Binghong Chen, Tianzhe Wang, Chengtao Li, Hanjun Dai, Le Song
    International Conference on Learning Representations (ICLR) 2021.
    [ Paper ] [ Code ]

  • BUSTLE: Bottom-Up Program Synthesis Through Learning-Guided Exploration
    Augustus Odena*, Kensen Shi*, David Bieber, Rishabh Singh, Charles Sutton, Hanjun Dai
    International Conference on Learning Representations (ICLR) 2021. Spotlight
    [ Paper ]

  • Learning Discrete Energy-based Models via Auxiliary-variable Local Exploration
    Hanjun Dai, Rishabh Singh, Bo Dai, Charles Sutton, Dale Schuurmans
    Advances in Neural Information Processing Systems (NeurIPS) 2020.
    [ arxiv ] [ Code ]

  • Differentiable Top-k Operator with Optimal Transport
    Yujia Xie, Hanjun Dai, Minshuo Chen, Bo Dai, Tuo Zhao, Hongyuan Zha, Wei Wei, Tomas Pfister
    Advances in Neural Information Processing Systems (NeurIPS) 2020.
    [ arxiv ] [ Code ]

  • Scalable Deep Generative Modeling for Sparse Graphs
    Hanjun Dai, Azade Nazi, Yujia Li, Bo Dai, Dale Schuurmans
    International Conference on Machine Learning (ICML) 2020.
    [ arxiv ] [ Code ]

  • Energy-Based Processes for Exchangeable Data
    Mengjiao Yang*, Bo Dai*, Hanjun Dai, Dale Schuurmans
    International Conference on Machine Learning (ICML) 2020.
    [ arxiv ] [ Code ]

  • Learning To Stop While Learning To Predict
    Xinshi Chen, Hanjun Dai, Yu Li, Xin Gao, Le Song
    International Conference on Machine Learning (ICML) 2020.
    [ arxiv ] [ Code ]

  • Retro*: Learning Retrosynthetic Planning with Neural Guided A* Search
    Binghong Chen, Chengtao Li, Hanjun Dai, Le Song
    International Conference on Machine Learning (ICML) 2020.
    [ arxiv ] [ Code ]

  • Code2Inv: A Deep Learning Framework for Program Verification
    Xujie Si*, Aaditya Naik*, Hanjun Dai, Mayur Naik and Le Song.
    International Conference on Computer-Aided Verification (CAV) 2020.
    [ Paper ] [ Project page ]

  • Hoppity: Learning Graph Transformations to Detect and Fix Bugs in Programs
    Elizabeth Dinella*, Hanjun Dai*, Ziyang Li, Mayur Naik, Le Song and Ke Wang.
    International Conference on Learning Representations (ICLR) 2020. Spotlight
    [ Paper ] [ Code ]

  • Retrosynthesis Prediction with Conditional Graph Logic Network
    Hanjun Dai, Chengtao Li, Connor Coley, Bo Dai and Le Song.
    Advances in Neural Information Processing Systems (NeurIPS) 2019
    [ arxiv ] [ Code ]

  • Learning Transferable Graph Exploration
    Hanjun Dai, Yujia Li, Chenglong Wang, Rishabh Singh, Po-Sen Huang and Pushmeet Kohli.
    Advances in Neural Information Processing Systems (NeurIPS) 2019
    [ arxiv ]

  • Exponential Family Estimation via Adversarial Dynamics Embedding
    Bo Dai, Zhen Liu, Hanjun Dai, Niao He, Arthur Gretton, Le Song and Dale Schuurmans.
    Advances in Neural Information Processing Systems (NeurIPS) 2019
    [ Paper ] [ arxiv ] [ Code ]

  • Particle Flow Bayes’ Rule
    Xinshi Chen, Hanjun Dai and Le Song.
    International Conference on Machine Learning (ICML) 2019
    [ arxiv ] [ Code ]

  • CompILE: Compositional Imitation Learning and Execution
    Thomas Kipf, Yujia Li, Hanjun Dai, Vinicius Zambaldi, Alvaro Sanchez-Gonzalez, Edward Grefenstette, Pushmeet Kohli and Peter Battaglia.
    International Conference on Machine Learning (ICML) 2019
    [ arxiv ] [ Code ]

  • Kernel Exponential Family Estimation via Doubly Dual Embedding
    Bo Dai*, Hanjun Dai*, Arthur Gretton, Le Song, Dale Schuurmans, Niao He (*Equal contributions).
    International Conference on Artificial Intelligence and Statistics (AISTATS) 2019,
    [ arxiv ] [ Code ]

  • Learning a Meta-Solver for Syntax-Guided Program Synthesis
    Xujie Si*, Yuan Yang*, Hanjun Dai, Mayur Naik, Le Song (*Equal contributions).
    International Conference on Learning Representations (ICLR) 2019,
    [ Paper ] [ Code ]

  • Learning Loop Invariants for Program Verification
    Xujie Si*, Hanjun Dai*, Mukund Raghothaman, Mayur Naik and Le Song (*Equal contributions).
    Advances in Neural Information Processing Systems (NIPS) 2018, Spotlight
    [ Paper ] [ Code ]

  • Coupled Variational Bayes via Optimization Embedding
    Bo Dai*, Hanjun Dai*, Niao He, Weiyang Liu, Zhen Liu, Jianshu Chen, Lin Xiao and Le Song (*Equal contributions).
    Advances in Neural Information Processing Systems (NIPS) 2018
    [ Paper ] [ Code ]

  • Cooperative neural networks (CoNN): Exploiting prior independence structure for improved classification
    Harsh Shrivastava, Eugene Bart, Bob Price, Hanjun Dai, Bo Dai and Srinivas Aluru.
    Advances in Neural Information Processing Systems (NIPS) 2018
    [ Paper ] [ Code ]

  • Learning Steady-States of Iterative Algorithms over Graphs
    Hanjun Dai, Zornitsa Kozareva, Bo Dai, Alexander Smola and Le Song.
    International Conference on Machine Learning (ICML) 2018
    [ Paper ] [ Code ]

  • Adversarial Attack on Graph Structured Data
    Hanjun Dai, Hui Li, Tian Tian, Xin Huang, Lin Wang, Jun Zhu, and Le Song.
    International Conference on Machine Learning (ICML) 2018
    [ arxiv ] [ Code ]

  • Syntax-Directed Variational Autoencoder for Structured Data.
    Hanjun Dai*, Yingtao Tian*, Bo Dai, Steven Skiena and Le Song (*Equal contributions).
    International Conference on Learning Representations (ICLR) 2018
    [ arxiv ] [ Code ]

  • Variational Reasoning for Question Answering with Knowledge Graph.
    Yuyu Zhang*, Hanjun Dai*, Zornitsa Kozareva, Alexander Smola and Le Song (*Equal contributions).
    AAAI Conference on Artificial Intelligence (AAAI) 2018. Oral
    [ arxiv ] [ Code ]

  • Learning Combinatorial Optimization Algorithms over Graphs
    Hanjun Dai*, Elias B. Khalil*, Yuyu Zhang, Bistra Dilkina and Le Song (*Equal contributions).
    Advances in Neural Information Processing Systems (NIPS) 2017. Spotlight
    [ arxiv ] [ Code ]

  • Know-Evolve: Deep Temporal Reasoning for Dynamic Knowledge Graphs
    Rakshit Trivedi, Hanjun Dai, Yichen Wang and Le Song.
    International Conference on Machine Learning (ICML) 2017
    [ arxiv ] [ Code ]

  • Recurrent Hidden Semi-Markov Model.
    Hanjun Dai, Bo Dai, Yan-Ming Zhang, Shuang Li and Le Song.
    International Conference on Learning Representations (ICLR) 2017
    [ Paper ] [ Code ]

  • Recurrent Coevolutionary Feature Embedding Processes for Recommendation
    Hanjun Dai*, Yichen Wang*, Rakshit Trivedi and Le Song (*Equal contributions)
    Recsys Workshop on Deep Learning for Recommender Systems (DLRS), 2016. Best Paper
    [ Paper ] [ Code ]

  • Recurrent Marked Temporal Point Processes: Embedding Event History to Vector.
    Nan Du, Hanjun Dai, Rakshit Trivedi, Utkarsh Upadhyay, Manuel Gomez-Rodriguez and Le Song.
    ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD) 2016.
    [ Paper ] [ Code ]

  • Discriminateive Embeddings of Latent Variable Models for Structured Data.
    Hanjun Dai, Bo Dai and Le Song.
    International Conference on Machine Learning (ICML) 2016.
    [ arxiv ] [ Code ]

  • Provable Bayesian Inference via Particle Mirror Descent.
    Bo Dai, Niao He, Hanjun Dai and Le Song.
    International Conference on Artificial Intelligence and Statistics (AISTATS), 2016. Best Student Paper
    [ Paper ]

  • M-Statistic for Kernel Change-Point Detection.
    Shuang Li, Yao Xie, Hanjun Dai and Le Song.
    Neural Information Processing Systems (NIPS), 2015.
    [ Paper ]

  • Sequential Click Prediction for Sponsored Search with Recurrent Neural Networks.
    Yuyu Zhang, Hanjun Dai, Chang Xu, Taifeng Wang, Jiang Bian and Tie-Yan Liu.
    AAAI Conference on Artificial Intelligence (AAAI), 2014.
    [ Paper ]

  • A Scalable Probabilistic Model for Learning Multi-Prototype Word Embedding
    Fei Tian, Hanjun Dai, Jiang Bian, Bin Gao, Rui Zhang and Tie-Yan Liu.
    International Conference on Computational Linguistics (COLING), 2014.
    [ Paper ]

Journal

  • Deep Coevolutionary Network: A Generic Embedding Framework for Temporally Evolving Graphs
    Hanjun Dai*, Yichen Wang*, Rakshit Trivedi and Le Song (*Equal contributions).
    [ arxiv ] [ Code ]

  • Material Structure-property Linkages Using Three-dimensional Convolutional Neural Networks.
    Ahmet Cecen, Hanjun Dai, Yuksel C. Yabansu, Surya R. Kalidindi and Le Song.
    Acta Materialia, 2017
    [ Paper ]

  • Sequence2Vec: A novel embedding approach for modeling transcription factor binding affinity landscape.
    Hanjun Dai*, Ramzan Umarov*, Hiroyuki Kuwahara, Yu Li, Le Song and Xin Gao(*Equal contributions).
    Bioinformatics, 2017, 1-9, DOI: 10.1093/bioinformatics/btx480
    [ Paper ] [ Supplement ] [ Code ]

  • KNET: A General Framework for Learning Word Embedding Using Morphological Knowledge.
    Qing Cui, Bin Gao, Jiang Bian, Siyu Qiu, Hanjun Dai and Tie-Yan Liu.
    ACM Transactions on Information Systems (TOIS), 2015.
    [ Paper ]

Workshop

  • Syntax-Directed Variational Autoencoder for Molecule Generation
    Hanjun Dai*, Yingtao Tian*, Bo Dai, Steven Skiena and Le Song (*Equal contributions)
    NIPS 2017 Workshop on Machine Learning for Molecules and Materials, 2017. Best Paper
    [ Paper ]

Software

GraphNN a general purpose deep neural network library with special design for structured data and dynamic computational graph. It comes with the unified CPU/GPU API. With the low-level support from Intel MKL and Cuda, it is very efficient.

[ Code][ Documentation]

Selected Awards

1st place in ByteCup International Machine Learning Competition 2016

Ranked 15 in ACM-ICPC World Finals 2015

Runner-up in ACM-ICPC 2014 Southeast USA Reginal and ACM-ICPC 2010 Dhaka Site

Chun-Tsung Scholar (established by Nobel Prize laureate, Tsung-Dao Lee)

CSC-IBM Excellence Scholarship 2013

National Scholarship, Fudan University, 2012

Activities

Invited Talks

Improved Generative Modeling of Structured Data.
RIKEN AIP Public , Tokyo (online), March 2021.
[ Event ]

Adversarial Attack on Graph Structured Data.
Cybersecurity Lecture Series (IISP, Gatech), Atlanta, March 2019.
[ Video ]

Learning with Structured Data.
Benevolent AI, London, Nov 2018.
[ Slides ]

Syntax-directed Variational Autoencoder for Molecule Generation.
Machine Learning for Molecules and Materials (NIPS Workshop Spotlight), Long Beach, Dec 2017.
[ pdf ]

Learning Combinatorial Optimization Algorithms over Graphs.
HotCSE Seminar (HotCSE), Atlanta, Nov 2017.
[ Abstract ] [ Slides ]

Graph Representation Learning with Deep Embedding Approach.
Machine Learning Conference (MLConf), Atlanta, Sep 2017.
[ Abstract ] [ Reference ] [ Video ] [ Slides ]

Teaching

Summer 2019, CS 3510, Design and analysis of algorithms

Spring 2019, CX 4240, Introduction to Computational Data Analysis

Fall 2015, CSE 6740, Computational Data Analysis

2015, Coach Assistant in Georgia Tech Programming Team [Link]

Service

Senior PC in AAAI 2019-2020, Reviewer in: NeurIPS 2016-2020, ICML 2017-2020, UAI 2020, AISTATS 2019, ICLR 2019-2020, Pattern Recognition 2018, KDD 2018-2020, AAAI 2018, KDD 2017, TKDD 2017

ICML 2016 Volunteer


The webpage template is kindly provided by Nan Du and Yingyu Liang