Stefano Favaro

Professor of Statistics, Università di Torino

Carlo Alberto Chair of Statistics and Machine Learning, Collegio Carlo Alberto

Departimento di Scienze Economico-Sociali e Matematico-Statistiche
Università di Torino
Corso Unione Sovietica 218/bis
10134 Torino, Italy

Email: stefano[.]favaro[at]unito[.]it

Cv [PDF]

Research interests: nonparametric Bayes and empirical Bayes methods, statistical machine learning, data confidentiality and fairness, learning-augmented recovery algorithms, mathematics of deep learning

Papers:

Main publications

Random measure priors in Bayesian recovery from sketches, with M. Beraha and M. Sesia
Journal of Machine Learning Research, 2024
[JMLR] [ArXiv]

MCMC for Bayesian nonparametric mixture modeling under differential privacy, with M. Beraha and V. Rao
Journal of Computational and Graphical Statistics, to appear
[JCGS] [ArXiv]

A martingale approach to Gaussian fluctuations and laws of iterated logarithm for Ewens-Pitman model, with B. Bercu
Stochastic Processes and their Applications, 2024
[SPA] [ArXiv]

Bayesian nonparametric inference for "species-sampling" problems, with C. Balocchi and Z. Naulet
Statistical Science, to appear
[STS] [ArXiv]

On second-order Poincaré inequalities in non-asymptotic approximation of Gaussian neural networks, with A. Bordino and S. Fortini
Advances in Approximate Bayesian Inference, 2024
[AABI]

Wasserstein posterior contraction rates in non-dominated Bayesian nonparametric models, with E. Dolera and E. Mainini
Annales de l'Institut Henri Poincaré - Probabilités et Statistiques, to appear
[AIHP] [ArXiv]

Scaled process priors for Bayesian nonparametric estimation of the unseen genetic variation, with T. Broderick, F. Camerlenghi and L. Masoero
Journal of the American Statistical Association, 2024
[JASA] [ArXiv]

Strong posterior contraction rates via Wasserstein dynamics, with E. Dolera and E. Mainini
Probability Theory and Related Fields, 2024
[PTRF] [ArXiv]

A Bayesian nonparametric approach to species sampling problems with ordering, with C. Balocchi and F. Camerlenghi
Bayesian Analysis, to appear
[BA] [ArXiv]

Bayesian nonparametric mixture modeling for temporal dynamics of gender stereotypes, with M. De Iorio, A. Guglielmi and Y. Lifeng
Annals of Applied Statistics, 2023
[AOAS] [ArXiv]

Near-optimal estimation of the unseen under regularly varying tail populations, with Z. Naulet
Bernoulli, 2023
[BEJ] [ArXiv]

Conformal frequency estimation using discrete sketched data with coverage for distinct queries, with E. Dobriban and M. Sesia
Journal of Machine Learning Research, 2023
[JMLR] [ArXiv]

Deep Stable neural networks: large-width asymptotics and convergence rates, with S. Fortini and S. Peluchetti
Bernoulli, 2023
[BEJ] [ArXiv]

Learning-augmented count-min sketches via Bayesian nonparametrics, with E. Dolera and S. Peluchetti
Journal of Machine Learning Research, 2023
[JMLR] [ArXiv]

Conformalized frequency estimation from sketched data, with M. Sesia
Neural Information Processing Systems, 2022
[NeurIPS]

Infinitely wide limits for deep Stable neural networks: sub-linear, linear and super-linear activation functions, with A. Bordino and S. Fortini
Transactions on Machine Learning Research, 2022
[TMLR] [ArXiv]

More for less: predicting and maximizing genetic variant discovery via Bayesian nonparametrics, with T. Broderick, F. Camerlenghi and L. Masoero
Biometrika, 2022
[Biometrika] [ArXiv]

Infinite-channel deep Stable convolutional neural networks, with D. Bracale, S. Fortini and S. Peluchetti
Neural Information Processing Systems, 2021
[NeurIPS]

Consistent and rate optimal estimation of the missing mass, with F. Ayed, M. Battiston and F. Camerlenghi
Annales de l'Institut Henri Poincaré - Probabilités et Statistiques, 2021
[AIHP] [ArXiv]

Bayesian nonparametric disclosure risk assessment, with F. Panero and T. Rigon
Electronic Journal of Statistics, 2021
[EJS]

Doubly infinite neural networks: a diffusion process approach, with S. Peluchetti
Journal of Machine Learning Research, 2021
[JMLR] [ArXiv]

Large-width functional asymptotics for deep Gaussian neural networks, with D. Bracale, S. Fortini and S. Peluchetti
International Conference on Learning Representations, 2021
[ICLR]

A Bayesian nonparametric approach to count-min sketch under power-law data streams, with E. Dolera and S. Peluchetti
Artificial Intelligence and Statistics, 2021
[AISTATS]

Optimal disclosure risk assessment, with F. Camerlenghi, Z. Naulet and F. Panero
Annals of Statistics, 2021
[AOS] [ArXiv]

Upscaling human activity data: an ecological perspective, with M. Formentin, A. Maritan, S. Stivanello, S. Suweis and A. Tovo
Proceedings of the National Academy of Sciences, 2021
[PNAS] [ArXiv]

Perfect sampling for posterior hierarchical Pitman-Yor processes, with S. Bacallado and L. Trippa
Bayesian Analysis, 2021
[BA]

Consistent estimation of small masses in feature sampling, with F. Ayed, M. Battiston and F. Camerlenghi
Journal of Machine Learning Research, 2021
[JMLR] [ArXiv]

Stable behaviour of infinitely wide deep neural networks, with S. Fortini and S. Peluchetti
Artificial Intelligence and Statistics, 2020
[AISTATS]

Infinitely deep neural networks as diffusion processes, with S. Peluchetti
Artificial Intelligence and Statistics, 2020
[AISTATS]

Nonparametric Bayesian multi-armed bandits for single cell experiment design, with F. Camerlenghi, B. Dimitrascu, B. Engelhardt and F. Ferrari
Annals of Applied Statistics, 2020
[AOAS] [ArXiv]

Rates of convergence in de Finetti's representation theorem, and Hausdorff moment problem, with E. Dolera
Bernoulli, 2020
[BEJ] [ArXiv]

A Berry-Esseen theorem for Pitman's alpha-diversity, with E. Dolera
Annals of Applied Probability, 2020
[AOAP] [ArXiv]

Approximating predictive probabilities of Gibbs-type priors, with J. Arbel
Sankhya Series A, 2020
[Sankhya] [ArXiv]

Bayesian mixed effects models for zero-inflated compositions in microbiome data analysis, with S. Bacallado, C. Huttenhower, B. Ren and L. Trippa
Annals of Applied Statistics, 2020
[AOAS] [ArXiv]

Neural SDE: information propagation through the lens of diffusion processes, with S. Peluchetti
Neural Information Processing Systems, 2019
[NeurIPS]

A Good-Turing estimator for feature allocation models, with F. Ayed, M. Battiston and F. Camerlenghi
Electronic Journal of Statistics, 2019
[EJS] [ArXiv]

Bayesian nonparametric analysis of Kingman's coalescent, with S. Feng and P. Jenkins
Annales de l'Institut Henri Poincaré - Probabilités et Statistiques, 2019
[AIHP] [ArXiv]

Modeling population structure under hierarchical Dirichlet processes, with K. Adhikari, M. De Iorio, L. Elliott and Y.W. Teh
Bayesian Analysis, 2019
[BA] [ArXiv]

Genomic variety prediction via Bayesian nonparametrics, with T. Broderick, F. Camerlenghi and L. Masoero
Advances in Approximate Bayesian Inference, 2019
[AABI]

Multi-armed bandits for species discovery: a Bayesian nonparametric approach, with M. Battiston and Y.W. Teh
Journal of the American Statistical Association, 2018
[JASA]

Dependent generalized Dirichlet priors for the analysis of acute lymphoblastic leukaemia, with W. Barcella, M. De Iorio and G. Rosner
Biostatistics, 2018
[Biostatistics]

Moderate deviations for Ewens-Pitman sampling models, with S. Feng and F. Gao
Sankhya Series A, 2018
[Sankhya] [ArXiv]

A characterization of product-form exchangeable feature probability functions, with M. Battiston, D.M. Roy and Y.W. Teh
Annals of Applied Probability, 2018
[AOAP] [ArXiv]

Posterior representations of hierarchical completely random measures in trait allocation models, with T. Broderick, F. Camerlenghi and L. Masoero
Neural Information Processing Systems, 2018
[NeurIPS]

On a general Maclaurin's inequality, with S.G. Walker
Proceedings of the American Mathematical Society, 2018
[PAMS]

Sufficientness postulates for Gibbs-type priors and hierarchial generalizations, with S. Bacallado, M. Battiston and L. Trippa
Statistical Science, 2017
[STS]

A marginal sampler for sigma-stable Poisson-Kingman mixture models, with M. Lomeli and Y.W. Teh
Journal of Computational and Graphical Statistics, 2017
[JCGS] [ArXiv]

Bayesian nonparametric inference for discovery probabilities: credible intervals and large sample asymptotics, with J. Arbel, B. Nipoti and Y.W. Teh
Statistica Sinica, 2017
[SS] [ArXiv]

Bayesian nonparametric ordination for the analysis of microbial communities, with S. Bacallado, S. Holmes, B. Ren and L. Trippa
Journal of the American Statistical Association, 2017
[JASA] [ArXiv]

Rediscovery of Good-Turing estimators via Bayesian nonparametrics, with B. Nipoti and Y.W. Teh
Biometrics, 2016
[Biometrics] [ArXiv]

Frequency of frequencies distributions and size-dependent exchangeable random partitions, with S.G. Walker and M. Zhou
Journal of the American Statistical Association, 2016
[JASA] [ArXiv]

On the stick-breaking representation for homogeneous NRMIs, with A. Lijoi, B. Nipoti, I. Pruenster and Y.W. Teh
Bayesian Analysis, 2016
[BA]

A note on nonparametric inference for species variety with Gibbs-type priors, with L.F. James
Electronic Journal of Statistics, 2016
[EJS]

Relatives of the Ewens sampling formula in Bayesian nonparametrics, with L.F. James
Statistical Science, 2016
[STS]

Large deviation principles for the Ewens-Pitman sampling model, with S. Feng
Electronic Journal of Probability, 2015
[EJP] [ArXiv]

A hybrid sampler for Poisson-Kingman mixture models, with M. Lomeli and Y.W. Teh
Neural Information Processing Systems, 2015
[NeurIPS]

Random variate generation for Laguerre-type exponentially tilted alpha-stable distributions, with B. Nipoti and Y.W. Teh
Electronic Journal of Statistics, 2015
[EJS]

Looking-backward probabilities for Gibbs-type exchangeable random partitions, with S. Bacallado and L. Trippa
Bernoulli, 2015
[BEJ] [ArXiv]

Bayesian regularization of the length of memory in reversible sequences, with S. Bacallado, L. Trippa
Journal of the Royal Statistical Society Series B, 2015
[JRSSB]

On the stick-breaking representation of sigma-stable Poisson-Kingman models, with M. Lomeli, B. Nipoti and Y.W. Teh
Electronic Journal of Statistics, 2014
[EJS]

Posterior analysis of rare variants in Gibbs-type species sampling models, with O. Cesari and B. Nipoti
Journal of Multivariate Analysis, 2014
[JMA]

Asymptotics for the number of blocks in a conditional Ewens-Pitman sampling model, with S. Feng
Electronic Journal of Probability, 2014
[EJP]

MCMC for normalized random measure mixture models, with Y.W. Teh
Statistical Science, 2013
[STS] [ArXiv]

Conditional formulae for Gibbs-type exchangeable random partitions, with A. Lijoi and I. Pruenster
Annals of Applied Probability, 2013
[AOAP] [ArXiv]

Slice sampling sigma-stable Poisson-Kingman mixture models, with S.G. Walker
Journal of Computational and Graphical Statistics, 2013
[JCGS]

Bayesian nonparametric analysis of reversible Markov chains, with S. Bacallado and L. Trippa
Annals of Statistics, 2013
[AOS] [ArXiv]

Asymptotics for a Bayesian nonparametric estimator of species richness, with A. Lijoi and I. Pruenster
Bernoulli, 2012
[BEJ] [ArXiv]

A class of normalized random measure with an exact predictive sampling scheme, with L. Trippa
Scandinavian Journal of Statistics, 2012
[SJS]

On the stick-breaking representation of normalized inverse Gaussian priors, with A. Lijoi and I. Pruenster
Biometrika, 2012
[Biometrika]

A new estimator of the discovery probability, with A. Lijoi and I. Pruenster
Biometrics, 2012
[Biometrics]

Alpha-diversity processes and normalized inverse Gaussian diffusions, with M. Ruggiero and S.G. Walker
Annals of Applied Probability, 2012
[AOAP] [ArXiv]

A class of measure-valued Markov chains and Bayesian nonparametrics, with A. Guglielmi and S.G. Walker
Bernoulli, 2012
[BEJ] [ArXiv]

On a class of random probability measures with a general predictive structure, with I. Pruenster and S.G. Walker
Scandinavian Journal of Statistics, 2011
[SJS]

Bayesian nonparametric inference for species variety with a two parameter Poisson-Dirichlet process prior, with A. Lijoi, R.H. Mena and I. Pruenster
Journal of the Royal Statistical Society Series B, 2009
[JRSSB]

A Gibbs-sampler based random process in Bayesian nonparametrics, with M. Ruggiero and S.G. Walker
Electronic Journal of Statistics, 2009
[EJS]

A generalized constructive definition for the Dirichlet process, with S.G. Walker
Statistics and Probability Letters, 2008
[SPL]

Other publications

On Johnson's "sufficientness" postulates for feature sampling models, with F. Camerlenghi and E. Dolera
Mathematics - A festschrift for Eugenio Regazzini's 75th birthday, 2021
[Mathematics] [ArXiv]

A compound Poisson perspective of Ewens-Pitman sampling model, with E. Dolera
Mathematics - A festschrift for Eugenio Regazzini's 75th birthday, 2021
[Mathematics] [ArXiv]

Discussion of "Sparse graphs using exchangeable random measures" by F. Caron and E.B. Fox, with M. Battiston
Journal of the Royal Statistical Society Series B, 2017
[JRSSB] [ArXiv]

Addendum to "On a general Maclaurin's inequality" with S.G.Walker
Proceedings of the American Mathematical Society, 2018
[PAMS]

Bayesian nonparametric inference for shared species richness in multiple populations, with S. Bacallado and L. Trippa
Journal of Statistical Planning and Inference - Special issue on Bayesian nonparametrics, 2015
[JSPI]

Bayesian inference on population structure: from parametric to nonparametric modeling, with M. De Iorio and Y.W. Teh
Nonparametric Bayesian Inference in Biostatistics, 2015
[Springer]

Are Gibbs-type priors the most natural generalization of the Dirichlet process?, with P. De Blasi, A. Lijoi, R.H. Mena, I. Pruenster and M. Ruggiero
IEEE Transactions on Pattern Analysis and Machine Intelligence - Special issue on Bayesian nonparametrics, 2015
[TPAMI] [ArXiv]

On a class of sigma-stable Poisson-Kingman models and an effective marginalized sampler, with M. Lomeli and Y.W. Teh
Statistics and Computing - Special issue MCMSki IV, 2015
[SC]

Discussion of "On simulation and properties of stable law" by L. Devroye and L.F. James, with B. Nipoti
Statistical Methods and Applications, 2014
[SMA]

Two tales about Bayesian nonparametric modeling, with P. De Blasi, A. Lijoi, R.H. Mena and I. Pruenster
Workshop on "Bayesian Statistical Sciences" at JSM, 2012
[JSM]

Contributions to the Dirichlet process and related classes of random probability measures
Ph.D. Thesis, Dipartimento di Scienze delle Decisioni, Università Commerciale "L. Bocconi", 2009
[Università "L. Bocconi"]

Submitted papers

Large-width asymptotics and training dynamics for alpha-Stable ReLU neural networks, with S. Fortini and S. Peluchetti
[ArXiv] 2024

The power of private likelihood-ratio tests for goodness-of-fit in frequency tables, with M. Beraha and E. Dolera
[ArXiv] 2024

Optimal estimation of high-order missing masses, and the rare-type match problem, with Z. Naulet
[ArXiv] 2024

A nonparametric Bayes approach to online activity prediction, with M. Beraha, L. Masoero, T. Richardson
[ArXiv] 2024

Quantitative central limit theorems in deep neural networks, with B. Hanin, D. Marinucci, I. Nourdin and G. Peccati
[ArXiv] 2024

A quasi-Bayesian sequential approach to deconvolution density estimation, with S. Fortini
[ArXiv] 2024

A smoothed-Bayesian approach to frequency recovery from sketched data, with M. Beraha and M. Sesia
[ArXiv] 2024

Function-space MCMC for Bayesian wide neural networks, with L. Pezzetti and S. Peluchetti
[ArXiv] 2024

Editorial boards: I am Associate Editor of Bernoulli and Statistical Science

Some links: