Jean Honorio

Senior Lecturer in the School of Computing and Information Systems at The University of Melbourne.
Adjunct Professor in the Computer Science Department at Purdue.
Room 3308, Melbourne Connect, Carlton, VIC 3053, Australia.
e-mail: jhonorio at unimelb.edu.au, jhonorio at purdue.edu, jhonorio at csail.mit.edu

Modern machine learning (ML) problems are combinatorial and non-convex, specially those involving latent variables, for which theoretical guarantees are quite limited. Furthermore, while quantitative guarantees (e.g., small test error) have been studied, qualitative guarantees (e.g., outlier robustness) are mostly lacking. My long-term research goal is to uncover the general foundations of ML and optimization that drives the empirical success across many specific combinatorial and non-convex ML problems. I aim to develop a set of optimization-theoretic frameworks and tools to bridge the aforementioned gaps, to further our understanding of continuous (possibly non-convex) relaxations of combinatorial problems, as well as our knowledge of non-convexity.

My aim is to generate correct, computationally efficient and statistically efficient algorithms for high dimensional ML problems. My research group has produced breakthroughs not only on classical worst-case NP-hard problems, such as learning and inference in structured prediction, community detection and learning Bayesian networks, but also on areas of recent interest such as fairness, meta learning, federated learning and robustness. [vita]

Prior to joining The University of Melbourne, I was Assistant Professor at Purdue and a postdoctoral associate at MIT CSAIL, working with Tommi Jaakkola. My Erdős number is 3: Jean Honorio → Tommi Jaakkola → Noga Alon → Paul Erdős.

Students

Current. Hanbyul Lee (Stat PhD), Wenjie Li (Stat PhD), Vineet Malik (CS MS), Deepak Maurya (CS PhD).
Past. Adarsh Barik (CS PhD 2023, postdoc at NUS), Chuyang Ke (CS PhD 2023), Kevin Bello (CS PhD 2021, postdoc at UChicago & CMU), Asish Ghoshal (CS PhD 2019).
Other co-authors. Imon Banerjee (Stat PhD), Huiming Xie (Stat PhD).
Other past co-authors. Donald Adams (CS BS), Site Bai (CS PhD), Gregory Dexter (CS PhD), Abdulrahman Alabdulkareem (CS BS), Longyun Guo (Ag PhD), Abi Komanduru (Eng PhD), Krishna Kesari (Eng MS), Yu-Jun Li (CS MS), Zitao Li (CS PhD), Jiajun Liang (Stat MS), Meimei Liu (Stat PhD), Raphael Meyer (CS BS), Yuki Ohnishi (Stat PhD), Keehwan Park (CS MS), Zhanyu Wang (Stat PhD), Zhaosen Wang (CS MS), Qiuling Xu (CS PhD), Yixi Xu (Stat PhD), Xiaochen Yang (Stat PhD), Qian Zhang (Stat MS), Yilin Zheng (CS MS).
Prospective. Here is a note for students who are considering working with me.

All Publications

Manuscripts under submission

Matrix Completion from General Deterministic Sampling Patterns. (Preprint)
Lee H., Mazumder R., Song Q., Honorio J.
(Under submission.)

Invex Programs: First Order Algorithms and Their Convergence. (Preprint)
Barik A., Sra S., Honorio J.
(Under submission.)

Minimax Optimal Rates For Robust GLMs Against Constant Proportion of Identifiable Outliers.
Maurya D., Honorio J.
(Under submission.)

On the True Geometry and Sample Complexity of Outlier-Robust Sparse Precision Matrix Estimation (Preprint)
Maurya D., Honorio J.
(Under submission.)

Outlier-Robust Estimation of a Sparse Linear Model Using Invexity. (Preprint)
Barik A., Honorio J.
(Under submission.)

Partial Inference in Structured Prediction. (Preprint)
Ke C., Honorio J.
(Under submission.)

A Novel Plug-and-Play Approach for Adversarially Robust Generalization. (Preprint)
Maurya D., Barik A., Honorio J.
(Under submission.)

A Theoretical Study of the Effects of Adversarial Attacks on Sparse Regression. (Preprint)
Maurya D., Honorio J.
(Under submission.)

Meta Sparse Principal Component Analysis. (Preprint)
Banerjee I., Honorio J.
(Under submission.)

Exact Inference with Latent Variables in an Arbitrary Domain. (Preprint)
Ke C., Honorio J.
(Under submission.)

2024

PyXAB - A Python Library for X-Armed Bandit and Online Blackbox Optimization Algorithms.
Li W., Li H., Song Q., Honorio J.
Journal of Open Source Software. [code]

An SDP Formulation for Minimizing p-th Order Controversy with Unknown Initial Opinions.
Chaitanya M., Barik A., Honorio J.
Complex Networks and Their Applications. Istanbul/Turkey.

Meta Learning for High-dimensional Ising Model Selection Using ℓ₁-regularized Logistic Regression.
Xie H., Honorio J.
Transactions on Machine Learning Research.

Support Recovery in Sparse PCA with General Missing Data.
Lee H., Song Q., Honorio J.
Uncertainty in Artificial Intelligence. Barcelona/Spain.

Identifying Causal Changes Between Linear Structural Equation Models.
Malik V., Bello K., Ghoshal A., Honorio J.
Uncertainty in Artificial Intelligence. Barcelona/Spain.

Personalized Federated X-Armed Bandit.
Li W., Honorio J., Song Q.
Artificial Intelligence and Statistics. Valencia/Spain.

On the Dual Problem of Convexified Convolutional Neural Networks.
Bai S., Ke C., Honorio J.
Transactions on Machine Learning Research.

Provable Guarantees for Sparsity Recovery with Deterministic Missing Data Patterns.
Ke C., Honorio J.
Transactions on Machine Learning Research.

Federated X-Armed Bandit.
Li W., Song Q., Honorio J., Lin G.
Association for the Advancement of Artificial Intelligence Conference. Vancouver/Canada.

Recovering Exact Support in Federated lasso without Optimization.
Barik A., Honorio J.
Transactions on Machine Learning Research.

2023

Exact Inference in High-order Structured Prediction.
Ke C., Honorio J.
International Conference on Machine Learning. Hawaii.

Remove Model Backdoors via Importance Driven Cloning.
Xu Q., Tao G., Honorio J., Liu Y., An S., Shen G., Cheng S., Zhang X.
IEEE Conference on Computer Vision and Pattern Recognition. Vancouver/Canada.

Provable Computational and Statistical Guarantees for Efficient Learning of Continuous-Action Graphical Games.
Barik A., Honorio J.
IEEE International Conference on Acoustics, Speech and Signal Processing. Rhodes Island/Greece.

2022

Support Recovery in Sparse PCA with Incomplete Data.
Lee H., Song Q., Honorio J.
Neural Information Processing Systems. Louisiana.

Exact Partitioning of High-order Models with a Novel Convex Tensor Cone Relaxation.
Ke C., Honorio J.
Journal of Machine Learning Research, 23(284): pp. 1-28

Sparse Mixed Linear Regression with Guarantees: Taming an Intractable Problem with Invex Relaxation.
Barik A., Honorio J.
International Conference on Machine Learning. Maryland.

A Simple Unified Framework for High Dimensional Bandit Problems.
Li W., Barik A., Honorio J.
International Conference on Machine Learning. Maryland.

On the Fundamental Limits of Exact Inference in Structured Prediction.
Lee H., Bello K., Honorio J.
IEEE International Symposium on Information Theory. Espoo/Finland.

Exact Partitioning of High-order Planted Models with a Tensor Nuclear Norm Constraint.
Ke C., Honorio J.
IEEE International Conference on Acoustics, Speech and Signal Processing. Singapore.

Provable Sample Complexity Guarantees for Learning of Continuous-Action Graphical Games with Nonparametric Utilities.
Barik A., Honorio J.
IEEE International Conference on Acoustics, Speech and Signal Processing. Singapore.

Information Theoretic Limits for Standard and One-Bit Compressed Sensing with Graph-Structured Sparsity.
Barik A., Honorio J.
IEEE International Conference on Acoustics, Speech and Signal Processing. Singapore.

A Thorough View of Exact Inference in Graphs from the Degree-4 Sum-of-Squares Hierarchy.
Bello K., Ke C., Honorio J.
Artificial Intelligence and Statistics. Virtual.

Federated Myopic Community Detection with One-shot Communication.
Ke C., Honorio J.
Artificial Intelligence and Statistics. Virtual.

2021

Fair Sparse Regression with Clustering: An Invex Relaxation for a Combinatorial Problem. (Spotlight)
Barik A., Honorio J.
Neural Information Processing Systems. Virtual.

Inverse Reinforcement Learning in a Continuous Space with Formal Guarantees.
Dexter G., Bello K., Honorio J.
Neural Information Processing Systems. Virtual.

A Lower Bound for the Sample Complexity of Inverse Reinforcement Learning.
Komanduru A., Honorio J.
International Conference on Machine Learning. Virtual.

Meta Learning for Support Recovery in High-dimensional Precision Matrix Estimation.
Zhang Q., Zheng Y., Honorio J.
International Conference on Machine Learning. Virtual.

A Le Cam Type Bound for Adversarial Learning and Applications.
Bello K., Xu Q., Honorio J.
IEEE International Symposium on Information Theory. Melbourne/Australia.

Information Theoretic Limits of Exact Recovery in Sub-hypergraph Models for Community Detection.
Liang J., Ke C., Honorio J.
IEEE International Symposium on Information Theory. Melbourne/Australia.

Information-Theoretic Bounds for Integral Estimation.
Adams D., Barik A., Honorio J.
IEEE International Symposium on Information Theory. Melbourne/Australia.

First Order Methods take Exponential Time to Converge to Global Minimizers of Non-Convex Functions.
Kesari K., Honorio J.
IEEE International Symposium on Information Theory. Melbourne/Australia.

Information-Theoretic Lower Bounds for Zero-Order Stochastic Gradient Estimation.
Alabdulkareem A., Honorio J.
IEEE International Symposium on Information Theory. Melbourne/Australia.

Regularized Loss Minimizers with Local Data Perturbation: Consistency and Data Irrecoverability.
Li Z., Honorio J.
IEEE International Symposium on Information Theory. Melbourne/Australia.

The Sample Complexity of Meta Sparse Regression.
Wang Z., Honorio J.
Artificial Intelligence and Statistics. California.

Novel Change of Measure Inequalities with Applications to PAC-Bayesian Bounds and Monte Carlo Estimation.
Ohnishi Y., Honorio J.
Artificial Intelligence and Statistics. California.

Randomized Deep Structured Prediction for Discourse-Level Processing.
Widmoser M., Pacheco M., Honorio J., Goldwasser D.
European Association for Computational Linguistics Conference. Kyiv/Ukraine.

PrivSyn: Differentially Private Data Synthesis.
Zhang Z., Wang T., Honorio J., Li N., Backes M., He S., Chen J. Zhang Y.
USENIX Security Symposium. Vancouver/Canada.

2020

Fairness Constraints can Help Exact Inference in Structured Prediction.
Bello K., Honorio J.
Neural Information Processing Systems. Vancouver/Canada.

Provable Efficient Skeleton Learning of Encodable Discrete Bayes Nets in Poly-Time and Sample Complexity.
Barik A., Honorio J.
IEEE International Symposium on Information Theory. California.

Minimax Bounds for Structured Prediction Based on Factor Graphs.
Bello K., Ghoshal A., Honorio J.
Artificial Intelligence and Statistics. Palermo/Italy.

Information Theoretic Sample Complexity Lower Bound for Feed-Forward Fully-Connected Deep Networks.
Yang X., Honorio J.
Technical report.

2019

Exact Inference in Structured Prediction.
Bello K., Honorio J.
Neural Information Processing Systems. Vancouver/Canada.

Learning Bayesian Networks with Low Rank Conditional Probability Tables.
Barik A., Honorio J.
Neural Information Processing Systems. Vancouver/Canada.

On the Correctness and Sample Complexity of Inverse Reinforcement Learning.
Komanduru A., Honorio J.
Neural Information Processing Systems. Vancouver/Canada.

Reconstructing a Bounded-Degree Directed Tree Using Path Queries.
Wang Z., Honorio J.
IEEE Allerton Conference on Communication, Control and Computing. Illinois.

Optimality Implies Kernel Sum Classifiers are Statistically Efficient.
Meyer R., Honorio J.
International Conference on Machine Learning. California.

Cost-Aware Learning for Improved Identifiability with Multiple Experiments.
Guo L., Honorio J., Morgan J.
IEEE International Symposium on Information Theory. Paris/France.

2018

Learning Latent Variable Structured Prediction Models with Gaussian Perturbations.
Bello K., Honorio J.
Neural Information Processing Systems. Montreal/Canada.

Computationally and Statistically Efficient Learning of Causal Bayes Nets Using Path Queries.
Bello K., Honorio J.
Neural Information Processing Systems. Montreal/Canada.

Information-Theoretic Limits for Community Detection in Network Models.
Ke C., Honorio J.
Neural Information Processing Systems. Montreal/Canada.

Statistically and Computationally Efficient Variance Estimator for Kernel Ridge Regression.
Liu M., Honorio J., Cheng G.
IEEE Allerton Conference on Communication, Control and Computing. Illinois.

Learning Maximum-A-Posteriori Perturbation Models for Structured Prediction in Polynomial Time. (Long presentation)
Ghoshal A., Honorio J.
International Conference on Machine Learning. Stockholm/Sweden.

Learning Linear Structural Equation Models in Polynomial Time and Sample Complexity.
Ghoshal A., Honorio J.
Artificial Intelligence and Statistics. Canary Islands/Spain.

Learning Sparse Polymatrix Games in Polynomial Time and Sample Complexity.
Ghoshal A., Honorio J.
Artificial Intelligence and Statistics. Canary Islands/Spain.

On the Statistical Efficiency of Compositional Nonparametric Prediction.
Xu Y., Honorio J., Wang X.
Artificial Intelligence and Statistics. Canary Islands/Spain.

The Error Probability of Random Fourier Features is Dimensionality Independent.
Li Y., Honorio J.
Technical report.

2017

Learning Identifiable Gaussian Bayesian Networks in Polynomial Time and Sample Complexity.
Ghoshal A., Honorio J.
Neural Information Processing Systems. California.

On the Sample Complexity of Learning Graphical Games.
Honorio J.
IEEE Allerton Conference on Communication, Control and Computing. Illinois.

Information Theoretic Limits for Linear Prediction with Graph-Structured Sparsity.
Barik A., Honorio J., Tawarmalani M.
IEEE International Symposium on Information Theory. Aachen/Germany.

Learning Graphical Games from Behavioral Data: Sufficient and Necessary Conditions.
Ghoshal A., Honorio J.
Artificial Intelligence and Statistics. Florida.

Information-Theoretic Limits of Bayesian Network Structure Learning.
Ghoshal A., Honorio J.
Artificial Intelligence and Statistics. Florida.

2016

From Behavior to Sparse Graphical Games: Efficient Recovery of Equilibria.
Ghoshal A., Honorio J.
IEEE Allerton Conference on Communication, Control and Computing. Illinois.

Structured Prediction: From Gaussian Perturbations to Linear-Time Principled Algorithms.
Honorio J., Jaakkola T.
Uncertainty in Artificial Intelligence. New York.

Information-Theoretic Lower Bounds for Recovery of Diffusion Network Structures.
Park K., Honorio J.
IEEE International Symposium on Information Theory. Barcelona/Spain.

Variable Selection in Gaussian Markov Random Fields.
Honorio J., Samaras D., Rish I., Cecchi G.
Invited book chapter in Log-Linear Models, Extensions and Applications.
Edited by Aravkin A., Deng L., Heigold G., Jebara T., Kanevski D., Wright S. (to be published on December, 2016)

2015

Learning the Structure and Parameters of Large-Population Graphical Games from Behavioral Data.
Honorio J., Ortiz L.
Journal of Machine Learning Research, 16(Jun): pp. 1157-1210 [code]

On the Statistical Efficiency of ℓ_1,p Multi-Task Learning of Gaussian Graphical Models.
Honorio J., Jaakkola T., Samaras D.
Technical report. [code]

Predictive Sparse Modeling of fMRI Data for Improved Classification, Regression, and Visualization Using the k-Support Norm.
Belilovsky E., Gkirtzou K., Misyrlis M., Konova A., Honorio J., Alia-Klein N., Goldstein R., Samaras D., Blaschko M.
Computerized Medical Imaging and Graphics, 46(1): pp. 40-46

Integration of PCA with a Novel Machine Learning Method for Reparameterization and Assisted History Matching Geologically Complex Reservoirs.
Honorio J., Chen C., Gao G., Du K., Jaakkola T.
Society of Petroleum Engineers: 91th Annual Technical Conference and Exhibition. Houston.

2014

A Unified Framework for Consistency of Regularized Loss Minimizers.
Honorio J., Jaakkola T.
International Conference on Machine Learning. Beijing/China.

Tight Bounds for the Expected Risk of Linear Classifiers and PAC-Bayes Finite-Sample Guarantees.
Honorio J., Jaakkola T.
Artificial Intelligence and Statistics. Reykjavik/Iceland.

Classification on Brain Functional Magnetic Resonance Imaging: Dimensionality, Sample Size, Subject Variability and Noise.
Honorio J.
Invited book chapter in Frontiers of Medical Imaging.
Edited by Chen C., World Scientific Publishing Company.

Improving Interpretability of Graphical Models in fMRI Analysis via Variable-Selection.
Honorio J., Samaras D., Rish I., Cecchi G.
Organization for Human Brain Mapping, Anual Meeting. Hamburg/Germany.

Predicting Cross-task Behavioral Variables from fMRI Data Using the k-Support Norm. (Best paper award)
Misyrlis M., Konova A., Blaschko M., Honorio J., Alia-Klein N., Goldstein R., Samaras D.
Medical Image Computing and Computer-Assisted Intervention. Workshop on Sparsity Techniques in Medical Imaging. Boston.

Methylphenidate Enhances Executive Function and Optimizes Prefrontal Function in Both Health and Cocaine Addiction.
Moeller S., Honorio J., Tomasi D., Parvaz M., Woicik P., Volkow N., Goldstein R.
Cerebral Cortex, 24(3): pp. 643-653

Integration of Principal Component Analysis and Streamline Information for the History Matching of Channelized Reservoirs.
Chen C., Gao G., Honorio J., Gelderblom P., Jimenez E., Jaakkola T.
Society of Petroleum Engineers: 90th Annual Technical Conference and Exhibition. Amsterdam/The Netherlands.

2013

Inverse Covariance Estimation for High-Dimensional Data in Linear Time and Space: Spectral Methods for Riccati and Sparse Models.
Honorio J., Jaakkola T.
Uncertainty in Artificial Intelligence. Washington. [code]

Two-Sided Exponential Concentration Bounds for Bayes Error Rate and Shannon Entropy.
Honorio J., Jaakkola T.
International Conference on Machine Learning. Atlanta.

fMRI Analysis of Cocaine Addiction Using k-Support Sparsity.
Gkirtzou K., Honorio J., Samaras D., Goldstein R., Blaschko M.
IEEE International Symposium on Biomedical Imaging. California. [code]

fMRI Analysis with Sparse Weisfeiler-Lehman Graph Statistics.
Gkirtzou K., Honorio J., Samaras D., Goldstein R., Blaschko M.
Medical Image Computing and Computer-Assisted Intervention, Workshop on Machine Learning in Medical Imaging. Nagoya/Japan. [code]

2012

Convergence Rates of Biased Stochastic Optimization for Learning Sparse Ising Models.
Honorio J.
International Conference on Machine Learning. Edinburgh/Scotland. [code]

Variable Selection for Gaussian Graphical Models.
Honorio J., Samaras D., Rish I., Cecchi G.
Artificial Intelligence and Statistics. Canary Islands/Spain. [code]

Can a Single Brain Region Predict a Disorder?
Honorio J., Tomasi D., Goldstein R., Leung H.C., Samaras D.
IEEE Transactions on Medical Imaging, 31(11): pp. 2062-2072 [code]

Two-person Interaction Detection Using Body-Pose Features and Multiple Instance Learning.
Yun K., Honorio J., Chattopadhyay D., Berg T., Samaras D.
IEEE Computer Vision and Pattern Recognition, Workshop on Human Activity Understanding from 3D Data. Rhode Island. [data]

Dopaminergic Involvement During Mental Fatigue in Health and Cocaine Addiction.
Moeller S., Tomasi D., Honorio J., Volkow N., Goldstein R.
Translational Psychiatry, 2: e176

Enhanced Midbrain Response at 6-month Follow-up in Cocaine Addiction, Association with Reduced Drug-related Choice.
Moeller S., Tomasi D., Woicik P., Maloney T., Alia-Klein N., Honorio J., Telang F., Wang G., Wang R., Sinha R., Carise D., Astone-Twerell J., Bolger J., Volkow N., Goldstein R.
Addiction Biology, 17(6): pp. 1013-25

2011

Lipschitz Parametrization of Probabilistic Graphical Models.
Honorio J.
Uncertainty in Artificial Intelligence. Barcelona/Spain.

Digital Analysis and Visualization of Swimming Motion.
Kirmizibayrak C., Honorio J., Jiang X., Mark R., Hahn J.
The International Journal of Virtual Reality, 10(3): pp. 9-16

Digital Analysis and Visualization of Swimming Motion.
Kirmizibayrak C., Honorio J., Jiang X., Mark R., Hahn J.
Conference on Computer Animation and Social Agents, Simulation of Sports Motion Workshop. Chengdu/China.

Dopaminergic contribution to endogenous motivation during cognitive control breakdown.
Moeller S., Tomasi D., Honorio J., Volkow N., Goldstein R.
Society for Neuroscience. Washington DC.

2010

Multi-Task Learning of Gaussian Graphical Models.
Honorio J., Samaras D.
International Conference on Machine Learning. Haifa/Israel. [code]

Simple Fully Automated Group Classification on Brain fMRI.
Honorio J., Samaras D., Tomasi D., Goldstein R.
IEEE International Symposium on Biomedical Imaging. Rotterdam/The Netherlands. [code]

Disrupted Functional Connectivity with Dopaminergic Midbrain in Cocaine Abusers.
Tomasi D., Volkow N., Wang R., Honorio J., Maloney T., Alia-Klein N., Woicik P., Telang F., Goldstein R.
Public Library of Science, PLoS ONE, 5(5): e10815

Oral Methylphenidate Normalizes Cingulate Activity in Cocaine Addiction During a Salient Cognitive Task.
Goldstein R., Woicik P., Maloney T., Tomasi D., Alia-Klein N., Shan J., Honorio J., Samaras D., Ruiliang W., Telang F., Wang G., Volkow N.
Proceedings of the National Academy of Sciences, 107(38): pp. 16667-72

2009

Sparse and Locally Constant Gaussian Graphical Models.
Honorio J., Ortiz L., Samaras D., Paragios N., Goldstein R.
Neural Information Processing Systems. Vancouver/Canada. [code]

Learning Brain fMRI Structure Through Sparseness and Local Constancy.
Honorio J., Ortiz L., Samaras D., Goldstein R.
Neural Information Processing Systems, Workshop on Connectivity Inference in NeuroImaging. Vancouver/Canada.

A Functional Geometry of fMRI BOLD Signal Interactions.
Langs G., Samaras D., Paragios N., Honorio J., Golland P., Alia-Klein N., Tomasi D., Volkow N., Goldstein R.
Neural Information Processing Systems, Workshop on Connectivity Inference in NeuroImaging. Vancouver/Canada.

Dopaminergic Response to Drug Words in Cocaine Addiction.
Goldstein R., Tomasi D., Alia-Klein N., Honorio J., Maloney T., Woicik P., Wang R., Telang F., Volkow N.
Journal of Neuroscience, 29(18): pp. 6001-6

Anterior Cingulate Cortex Hypoactivations to an Emotionally Salient Task in Cocaine Addiction.
Goldstein R., Alia-Klein N., Tomasi D., Honorio J., Maloney T., Woicik P., Wang R., Telang F., Volkow N.
Proceedings of the National Academy of Sciences, 106(23): pp. 9453-8

2008

Task-Specific Functional Brain Geometry from Model Maps.
Langs G., Samaras D., Paragios N., Honorio J., Alia-Klein N., Tomasi D., Volkow N., Goldstein R.
Medical Image Computing and Computer-Assisted Intervention. New York.