Rafael B. Stern

Professor of Statistics

Universidade de São Paulo

Biography

I am an Assistant Professor at the University of São Paulo. I am currently a member of the Scientific Council of the Brazilian Association of Jurimetrics, an associate investigator at NeuroMat and a member of the Order of Attorneys of Brazil. I am a member of the following research groups:

In case you are looking for Rafael Izbicki, his site is here.

Interests

• Bayesian Statistics
• Computational Historical Linguistics
• Decision Theory
• Jurimetrics
• Machine Learning

Education

• PhD in Statistics, 2015

  Carnegie Mellon University

• MSc in Statistics, 2008

  Universidade de São Paulo

• BA in Law, 2007

  Pontifícia Universidade Católica - SP

• BA in Statistics, 2006

  Universidade de São Paulo

Featured Publications

CD-split and HPD-split: Efficient conformal regions in high dimensions

Conformal methods create prediction bands that control average coverage assuming solely i.i.d. data. We introduce CD-split and HPD-split, which yield general prediction regions and converge to the optimal highest predictive density set.

Rafael Izbicki, Gilson Shimizu, Rafael B. Stern

PDF Cite Code

Indemnity for a lost chance

Civil liability for a lost chance occurs when a tortious action changes the probabilities of the outcomes that can be obtained by the victim. Instead of discussing specific formulae, we propose a principled approach to quantifying damages in lost chances. We present several qualitative legal questions regarding how to quantify damages and show that the answers lead to precise and general rules for damage quantification.

Rafael B. Stern, Joseph B. Kadane

PDF Cite Code Source Document DOI

Quantification under prior probability shift: the ratio estimator and its extensions

The quantification problem consists of determining the prevalence of a given label in a target population. However, one often has access to the labels in a sample from the training population but not in the target population. A common assumption in this situation is that of prior probability shift, that is, once the labels are known, the distribution of the features is the same in the training and target populations. In this paper, we derive a new lower bound for the risk of the quantification problem under the prior shift assumption. Using a weaker version of the prior shift assumption, which can be tested, we show that ratio estimators can be used to build confidence intervals for the quantification problem.

Afonso F. Vaz, Rafael Izbicki, Rafael B. Stern

PDF Cite

The Logical Consistency of Simultaneous Agnostic Hypothesis Tests

Simultaneous hypothesis tests can fail to provide results that meet logical requirements. For example, A implies B, there exist tests that, based on the same data, reject B but not A. This paper develops logically-coherent agnostic tests that also perform well statistically. These tests are characterized as region estimator-based tests.

Luís G. Esteves, Rafael Izbicki, Julio M. Stern, Rafael B. Stern

PDF Cite

Recent Publications

Luís G. Esteves, Rafael Izbicki, Julio M. Stern, Rafael B. Stern (2023). Logical coherence in Bayesian simultaneous three-way hypothesis tests. In IJAR.

PDF Cite DOI

Ruobin Gong, Joseph B.Kadane, Mark J. Schervish, Teddy Seidenfeld, Rafael B. Stern (2022). Learning and total evidence with imprecise probabilities. In IJAR.

PDF Cite DOI

Rafael Izbicki, Gilson Shimizu, Rafael B. Stern (2022). CD-split and HPD-split: Efficient conformal regions in high dimensions. In JMLR.

PDF Cite Code

Fabiana Luci de Oliveira, Rafael B. Stern, Ana Fernanda Noli (2022). Processo decisório do Supremo Tribunal Federal na transição democrática brasileira (1978-1988). In REED.

PDF Cite

Ruobin Gong, Joseph B. Kadane, Mark J. Schervish, Teddy Seidenfeld, Rafael B. Stern (2021). Deceptive Credences. In Ergo.

PDF Cite Source Document DOI

See all publications

Teaching

Undergraduate courses

• Bayesian Inference (16/1, 17/1, 21/1, 23/1)
• Causal Inference (22/2, 23/1)
• Classical Inference (17/2)
• Computational Statistics (21/1, 21/2)
• Introduction to Statistics (16/1, 18/1, 18/2, 19/1, 22/1, 22/2)
• Probability 1 (17/1, 18/2, 19/2)
• Probability 2 (16/2, 18/1)

Graduate courses

• Bayesian Inference (16/2, 19/1, 21/1, 21/2, 22/1)
• Causal Inference (23/1)
• Classical Inference (19/2)
• Decision Theory (17/2)

Recent & Upcoming Talks

Indemnity for a lost chance

Civil liability for a lost chance applies to cases in which a tortious action changes the probabilities of the outcomes that can be obtained by the victim. This talk uses Decision Theory to discuss how to quantify damages in such cases.

Jun 1, 2021 1:00 PM — 3:00 PM COMCA 2021

Rafael B. Stern, Joseph B. Kadane

Video

A model for the evolution of languages

I’ll present a model for infering the evolution of the phonology of languages. A relevant innovation of this model is that it captures the regularity of sound changes.

Jun 1, 2017 1:00 PM — 3:00 PM

Rafael B. Stern, Joseph B. Kadane

Video

Teoria da decisão e aplicações ao Direito

Introdução à Teoria da Decisão com aplicações na Jurimetria. Em particular, discute-se o efeito da ausência de taxas judiciais nos JEC’s sobre o valor pedido pelo autor.

Jun 1, 2017 1:00 PM — 3:00 PM UNICAMP

Rafael B. Stern

Video

Recent Posts

Doação de equipamento de proteção para hospitais

Descrição das atividades da Tropa da Garoa na produção e doação de equipamentos de proteção individual durante a pandemia COVID-19.

Rafael B. Stern

May 13, 2020 3 min read COVID-19, Tropa da Garoa

A discrepância entre o valor da causa e o valor concedido por danos morais nos JEC's

Análise baseada em Teoria da Decisão da discrepância entre o valor pedido pelo autor e o valor concedido pelo juiz nos Tribunais de pequenas causas (JEC).

Athos Damiani, Mariana Ferez, Rafael B. Stern, Julio A. Z. Trecenti

May 16, 2019 10 min read Jurimetrics

Jurimetrics in the Constitution

I recently wrote a text in portuguese which studies how Brazil’s Supreme Federal Court decided ex-president Lula’s habeas corpus. In the following, I translate this text to English. Brazil’s Supreme Federal Court recently decided ex-president Lula’s habeas corpus in a tight vote of 5 for and 6 against.

Rafael B. Stern

Apr 23, 2018 4 min read Jurimetrics

Students

Nicholas Moura e Silva	Graduate	2022	Aplicação de modelo de classificação em relatos de atendimento como método de automação de atividades jurídicas
Renato Annoni	Graduate	2022	Aferição da eficácia regulatóriado ambiente econômico financeiro via análise de dados
Humberto Bezerra de Meneses Júnior	Graduate	2022	Uma análise jurimétrica aplicada ao problema do anatocismo no Brasil
Gabriela Massoni	Graduate	2021	Análise de textos por meio de processos estocásticos na representação word2vec
Lucas Pereira Lopes	Graduate	2020	Aprendizado não-supervisionado utilizando representação em grafos
Elmo Henrique de Moraes	Graduate	2020	Classificação de contratos de seguros para atendimento da norma contábil
Guilherme Antonio Lima	Undergraduate	2020	Restauração de imagens e otimização por Monte Carlo
Luben Miguel Cabezas	Undergraduate	2020	Previsão em modelos filogenéticos
Luri Yasuda Ikeda	Undergraduate	2020	Métodos de simulação para o cálculo de integrais com aplicações à Inferência Bayesiana
Ana Fernanda Noli	Undergraduate	2020	Estimação de pontos ideais no STF em decisões de ADI e Rp
Ana Fernanda Noli	Undergraduate	2020	Estratégias ótimas em ações judiciais
Mayara Regina Formenton	Undergraduate	2020	Análise estatística do Judiciário no combate à violência contra a mulher
Tatyana Zabanova	Graduate	2019	Improving recommender systems with social network data
Joao Poloniato	Graduate	2018	FBST in high dimensions
Rafaela Vansan	Undergraduate	2018	Analysis of hamiltonian monte carlo
Carolina T. S. Mello	Undergraduate	2018	A map of administrative improbity in Brazil
Gabriela Massoni	Undergraduate	2018	The efficiency of shared judicial staff
Marcia Silva	Undergraduate	2017	Using police reports to determine the reason why a person is missing
Michelle Massaro	Undergraduate	2017	A model for child adoption matching

Contact

• rbstern at gmail dot com

Popular Topics

Agnostic Test Bernstein Polynomials Bets Causality Conformal Prediction Crowdsourcing Decision Theory Foundations Haphazard Sampling Hypothesis Test Intentional Sampling JEC Jurimetrics Logical Coherence Machine Learning Minimax Orgulity Prior Shift Quantification Rerandomization