(8/8) CONTROLLING THE FALSE DISCOVERY RATE VIA KNOCKOFFS

(8/29) Panning for gold:‘model-X’ knockoffs for high dimensional controlled variable selection

(9/12) ROBUST INFERENCE WITH KNOCKOFFS

(9/29) A KNOCKOFF FILTER FOR HIGH-DIMENSIONAL SELECTIVE INFERENCE

• 격주에 한 번, 책의 목차를 기준으로 약 2단원 분량을 공부하며, 필요한 경우 관련 논문을 함께 소개함.

• 발표자 : 김정주

• 책 : Hypothesis testing with e-values

• 발표자료 : [link]

• 각자 주어진 논문을 읽고 발표하는 방식으로 진행.

• 논문 목록

(8/2, 김정주) Testing for Outliers with Conformal p-values

(8/6, 서광옥) False Discovery Rate Control For Structured Multiple Testing

(8/12, 이슬) Integrative conformal p-values for powerful out-of-distribution testing with labeled outliers

(8/14, 신민섭) E-values: Calibration, combination, and applications

(8/16, 조성훈 교수) A Distribution-Free Empirical Bayes Approach to Multiple Testing with Side Information

• 2주 간 주 3회, 매일 2명의 발표자가 분배된 챕터를 발표하고 질의응답하는 방식으로 진행.

• 원중호 교수님 연구실과 함께 진행.

• 책 : Foundations of Quantization for Probability Distributions

• 발표자료 : [link]

• 각자 주어진 논문을 읽고 발표하는 방식으로 진행.

• 논문 목록

1. (서광옥) Knock off procedure

2. (김승규) DAG and variable ordination

3. (이다혜) variance estimation in high dimensional linear regression

4. (신민섭) Measurement error models: from nonparametric methods to deep neural networks

5. (신민섭) Bayesian Deep Net GLM and GLMM

(각자 읽어 볼 논문)

A Survey on Bayesian Deep Learning

A Selective Overview of Deep Learning

• 각자 주어진 논문을 읽고 발표하는 방식으로 진행.

• 일정

(3/12, 지승영) 1-1. Zhang, C., Chen, M., & Wang, X. (2020). Statistical methods for quantifying between-study heterogeneity in meta-analysis with focus on rare binary events. Statistics and Its Interface, 13(4), 449-464.

(3/12, 지승영) 1-2. Zhang, C., Wang, X., Chen, M. and Wang, T. (2021), A comparison of hypothesis tests for homogeneity in meta‐analysis with focus on rare binary events. Res Syn Meth. Accepted Author Manuscript.

(3/26, 조영현) 2. Ebadi, M., Chenouri, S. E., & Lin, D. K. (2020). Statistical monitoring of covariance matrix in multivariate processes: A literature review. arXiv preprint arXiv:2002.06159.

(4/9, 김승규) 3. Park, S. and Lim, J. (2021) An overview of heavy-tail extensions of multivariate gaussian distribution and their relationships.

• 발표자료 : [link]

Last update: 7월 2일

세미나 시각: 금요일 오전 10시, zoom meeting (id: TBA)

1. 7월 10일, 임창준,  Ji et al.2014, Differential Privacy and Machine Learning: a Survey and Review.

2. 7월 17일, 김영래, 

   • Dwork et al.2006, Calibrating noise to sensitivity in private data analysis.

   • Nissim et al.2007, Smooth Sensitivity and Sampling in Private Data Analysis.

3. 7월 24일, 박성오, Ren et al. 2018, LoPub: High-Dimensional Crowdsourced Data Publication With Local Differential Privacy

4. 7월 31일, 김승규,

   • DworkRothblum2016, Concentrated Differential Privacy

   • BunSteinke2016, Concentrated Differential Privacy: Simplifications, Extensions, and Lower Bounds

5. 8월 7일, 구태현, Asi et al.2019, Element Level Differential Privacy: The Right Granularity of Privacy

6. 8월 14일, 조영현, 

   • Charest2010: how to analyze differentially private data, 

   • McClureReiter2012: Differential Privacy and Statistical Disclosure Risk Measures: An Investigation with Binary Synthetic Data

7. 8월 21일, 이성진, Hardt et al. 2012, A Simple and Practical Algorithm for Differentially Private Data Release

8월 28일. 현형진, Jiang et al.2015, Wishart Mechanism for Differentially Private Principal Components Analysis SmithThakurta2013, Differentially Private Model Selection via Stability Arguments and the Robustness of the Lasso 

1. TBA, 조성훈, Amin et al.2019, Differentially Private Covariance Estimation

 Graphical model (Causal inference)

Summary by Namjoon Suh

1. Asymptotics of empirical eigenstructure for high dimensional spiked covariance (https://projecteuclid.org/euclid.aos/1497319697)

2. Asymptotic Theory of Eigenvectors for Large Random Matrices (https://arxiv.org/abs/1902.06846)

3. Summary of Yata's works (noise reduction, data transformation)

4. Bootstrapping spectral statistics in high dimensions (https://arxiv.org/abs/1709.08251)

5. Global testing against sparse alternatives under Ising models (https://projecteuclid.org/euclid.aos/1534492829)