List of Publications
For a list of code releases, see our research page.
Early Drafts/Preprints

Optimistic Estimation of Convergence in Markov Chains with the AverageMixing Time,
(Preprint) [arXiv] 
Dimensionfree Bounds for Sum of Dependent Matrices and Operators with HeavyTailed Distribution,
(Preprint) [arXiv] 
VarianceAware Estimation of Kernel Mean Embedding,
(Preprint) [arXiv] 
Optimal QuasiBayesian Reduced Rank Regression with Incomplete Response,
(Preprint) [arXiv] 
Concentration and Robustness of Discrepancybased ABC via Rademacher Complexity,
(Preprint) [arXiv]
In press / to appear

Geometric Aspects of DataProcessing of Markov Chains,
(Transactions of Mathematics and Its Applications) [arXiv]
2024

Variational Learning is Effective for Large Deep Networks,
(ICML 2024) [arXiv] [Blog] [Code]

Position Paper: Bayesian Deep Learning in the Age of LargeScale AI,
(ICML 2024) [arXiv]

Improved Estimation of Relaxation Time in Nonreversible Markov Chains,
(Annals of Applied Probability) [Published version] [arXiv]

Model Merging by UncertaintyBased Gradient Matching,
(ICLR 2024) [arXiv] [Code]

Conformal Prediction via RegressionasClassification,
(ICLR 2024) [OpenReview] [ArXiv] [Code] [Package]
2023

Improving Continual Learning by Accurate Gradient Reconstructions of the Past,
(TMLR) [ OpenReview ] [ Code ]

The Bayesian Learning Rule,
(JMLR) [ Published version ] [ arXiv ] [ Tweet ]

The Memory Perturbation Equation: Understanding Model’s Sensitivity to Data,
(NeurIPS 2023) [ arXiv ] [ SlidesLive Video ] [ Poster] [Code] 
Bridging the Gap Between Target Networks and Functional Regularization,
(TMLR) [ Openreview ] 
Variational Bayes Made Easy,
(AABI 2023) [arXiv]

Estimation of Copulas via Maximum Mean Discrepancy,
(JASA) [Journal version] [arXiv] 
Empirical and InstanceDependent Estimation of Markov Chain and Mixing Time,
(Scandinavian Journal of Statistics) [arXiv] [Journal version] 
Systematic Approaches to Generate Reversiblizations of Markov Chains,
(IEEE Transactions on Information Theory) [arXiv] [Early Access] 
Learning and Identity Testing of Markov Chains,
(Handbook of Statistics, Volume 49) [Journal version] 
Exploiting Inferential Structure in Neural Processes,
(UAI 2023) [Published version] [arXiv] [Poster] 
Information Geometry of Markov Kernels: a Survey,
in "Advances in Information Geometry: Beyond the Conventional Approach",
(Front. Phys. Sec. Statistical and Computational Physics) [Journal version] 
MasakhaPOS: PartofSpeech Tagging for Typologically Diverse African Languages,
(ACL 2023) [arXiv] 
Geometric Reduction for Identity Testing of Reversible Markov Chains,
(GSI 2023) [Published version] [arXiv]
Oral presentation. 
Simplifying Momentumbased Riemannian Submanifold Optimization,
(ICML 2023) [ ArXiv ]

MemoryBased Dual Gaussian Processes for Sequential Learning,
(ICML 2023) P. E. Chang, P. Verma, S. T. John, A. Solin, M.E. Khan</span> 
DimensionFree Empirical Entropy Estimation,
(IEEE Transactions on Information Theory) [Journal version] [arXiv] 
The LieGroup Bayesian Learning Rule,
(AISTATS 2023) [arXiv] [Code] 
SAM as an Optimal Relaxation of Bayes,
(ICLR 2023) [arXiv] [Code]
Notable top5% of all accepted papers.
2022

Sequential Learning in GPs with Memory and Bayesian Leverage Score,
(Continual Lifelong Workshop at ACML 2022) , [ OpenReview ]

MasakhaNER 2.0: Africacentric Transfer Learning for Named Entity Recognition,
(EMNLP 2022) [arXiv] 
Practical Structured Riemannian Optimization with Momentum by using Generalized Normal Coordinates,
(NeuReps Workshop at NeurIPS 2022) , [ OpenReview ]

Can Calibration Improve Sample Prioritization?,
(HITY Workshop at NeurIPS 2022) [ OpenReview ]

Exploiting Inferential Structure in Neural Processes,
(Workshop on Tractable Probabilistic Modeling at UAI 2022 ) [ OpenReview ] [Video] [Poster]

Deviation Inequalities for Stochastic Approximation by Averaging,
(SPA) [Published version] [arXiv] 
Understanding the Population Structure Correction Regression,
(ICSTA 2022) [Published version] [arXiv] 
Approximate Bayesian Inference: Reprint of the Special Issue Published in Entropy,
(MDPI Books) [Book page] 
Tight Risk Bound for High Dimensional Time Series Completion,
(EJS) [Published version] [arXiv] 
Finite Sample Properties of Parametric MMD Estimation: Robustness to Misspecification and Dependence,
(Bernoulli) [Published version] [arXiv]
2021

KnowledgeAdaptation Priors,
(NeurIPS 2021) [Published version] [arXiv] [Slides] [Tweet] [SlidesLive Video] [Code] 
Dual Parameterization of Sparse Variational Gaussian Processes,
(NeurIPS 2021) [Published version] [arXiv] [Code] 
Metastrategy for Learning Tuning Parameters with Guarantees,
(Entropy) [Published version] [arXiv] 
SubsetofData Variational Inference for Deep GaussianProcess Regression,
(UAI 2021) , [Published version] [arXiv] [Code] 
Scalable Marginal Likelihood Estimation for Model Selection in Deep Learning,
(ICML 2021) [Published version] [arXiv] [Code] 
Tractable Structured Natural Gradient Descent Using Local Parameterizations,
(ICML 2021) [Published version] [arXiv] 
NonExponentially Weighted Aggregation: Regret Bounds for Unbounded Loss Functions,
(ICML 2021) [Published version] [arXiv] 
Improving Predictions of Bayesian Neural Networks via Local Linearization,
(AIstats 2021) [Published version] [arXiv] [Code] 
A Theoretical Analysis of Catastrophic Forgetting through the NTK Overlap Matrix,
(AISTATS 2021) [Published version] [arXiv] 
Simultaneous Dimension Reduction and Clustering via the NMFEM Algorithm,
(Advances in Data Analysis and Classification) [Published version] [arXiv]
2020

Continual Deep Learning by Functional Regularisation of Memorable Past
(NeurIPS 2020) [Published version] [ArXiv] [Code] [Poster] Oral presentation, 1% of all submissions (105 out of 9454 submissions).

Approximate Bayesian Inference,
(Entropy) [Paper] 
Concentration of tempered posteriors and of their variational approximations,
(Annals of Statistics) [Published version] [arXiv] 
Fast Variational Learning in StateSpace Gaussian Process Models,
(MLSP 2020) [Published version] [arXiv] 
Highdimensional VAR with lowrank transition,
(Statistics and Computing) [Published version] [arXiv] 
AI for Social Good: Unlocking the Opportunity for Positive Impact,
(Nature Communications 2020)
[Paper] [Declaration on AI2SG] [Dagstuhl AI4SG 2019] [Press Release] 
Training Binary Neural Networks using the Bayesian Learning Rule,
(ICML 2020) [Published version] [arXiv] [Code] 
Handling the PositiveDefinite Constraint in the Bayesian Learning Rule,
(ICML 2020) [Published version] [arXiv] [Code] 
VILD: Variational Imitation Learning with Diversequality Demonstrations,
(ICML 2020) [Published version] [arXiv] [Code] 
MMDBayes: Bayesian Estimation via Maximum Mean Discrepancy,
(AABI 2019) [Published version] [arXiv] 
Exact Recovery of Lowrank Tensor Decomposition under Reshuffling,
(AAAI 2020) , [arXiv]
2019

Practical Deep Learning with Bayesian Principles,
(NeurIPS 2019) . [Published version] [arXiv] [Code] 
Approximate Inference Turns Deep Networks into Gaussian Processes,
(NeurIPS 2019) . [Published version] [arXiv] [Code] 
A Generalization Bound for Online Variational Inference (best paper award),
(ACML 2019) . [Published version] [arXiv] 
Matrix factorization for multivariate time series analysis,
(EJS) [Published version] [arXiv] 
Fast and Simple NaturalGradient Variational Inference with Mixture of Exponentialfamily Approximations,
(ICML, 2019) . [arXiv] [Published version] [Code]
Also appeared at the Symposium on Advances in Approximate Bayesian Inference at NeurIPS 2018 [Short Paper]

Stein's Lemma for the Reparameterization Trick with Exponential Family Mixtures,
(ICML workshop on Stein's Method in ML and Stats, 2019) . [arXiv] 
Scalable Training of Inference Networks for GaussianProcess Models,
(ICML, 2019) . [arXiv] [ Published version] [Code] 
TDRegularized ActorCritic Methods,
(Machine Learning, 2019. A short version appeared at EWRL 2018)
. [ Published version] [arXiv] [Short version at EWRL 2018]
2018

SLANG: Fast Structured Covariance Approximations for Bayesian Deep Learning with Natural Gradient,
(NeurIPS 2018) . [ Published version] [arXiv] [Poster] [3min Video] [Code] 
Natural Variational Continual Learning,
(Continual Learning Workshop at NIPS 2018)
[Paper]. 
Fast yet Simple NaturalGradient Descent for Variational Inference in Complex Models,
(ISITA 2018) , [arXiv] [IEEE explore] [Slides] 
Fast and Scalable Bayesian Deep Learning by WeightPerturbation in Adam,
(ICML 2018) , [ Published version] [arXiv] [Code] [Slides] 
Variational Message Passing with Structured Inference Networks,
(ICLR 2018) , [Paper] [arXiv Version] [Code] 
Bayesian Nonparametric PoissonProcess Allocation for TimeSequence Modeling,
(AIStats 2018) , , [Published version] [Code]
2017

Vprop: Variational Inference using RMSprop,
(NIPS 2017, Workshop on Bayesian Deep Learning)
[Workshop version] [Poster] 
Variational AdaptiveNewton Method for ExplorativeLearning,
(NIPS 2017, Workshop on Advances in Approximate Bayesian Inference)
[arXiv Version] [Poster] 
NaturalGradient Stochastic Variational Inference for NonConjugate Structured Variational Autoencoder,
(ICML 2017, Workshop on Deep Structure Prediction) , [Paper] 
ConjugateComputation Variational Inference : Converting Variational Inference in NonConjugate Models to Inferences in Conjugate Models,
(AIstats 2017) [ Published version ] [arXiv ] [Code for Logistic Reg + GPs] [Code for Correlated Topic Model] 
SmarPer: ContextAware and Automatic RuntimePermissions for Mobile Devices,
(38th IEEE Symposium on Security and Privacy (S&P;), San Jose, CA, USA, May 2224, 2017)
[Published paper] [Code] [SmarPer Homepage] 
GaussianProcessBased Emulators for Building Performance Simulation,
(Building Simulation 2017) [Paper] [Building Simulation Data] [Code]