9:00 a.m. | Welcome & Breakfast (Rm: Main Hall)
Conference check-in desk and light refreshments will be available in the FED Hall lobby.
9:15 a.m. | Opening Remarks
9:30 a.m. | Keynote (Rm: Main Hall)
Chair: Benjamin Côté
Dr. Mike Hoy, University of Guelph
Welfare Economics Meets Risk Theory
This seminar has three principal aims: (1) a short review of "standard" welfare economics (with some extras) that incorporates distributional concerns, (2) a discussion of how distributionally sensitive welfare analysis and inequality measurement have been influenced by risk theory concepts, (3) a few simple examples illustrating applications involving risk that help to connect theory to practice and also illustrate challenges in doing so.
10:30 a.m. | Coffee Break
Coffee will be served in the FED Hall Main Hall.
11:00 a.m. | Parallel Session 1
Session 1A (Rm: Columbia A)
Chair: Madison Cranstone
| Topics in Statistics | |
|---|---|
| Bryn Crandles | Accelerated Majorization-Minimization Algorithms for Estimating Common Eigenvector Matrices in Gaussian Mixture Models |
| Rohit Gajendragadkar | Comparing continuously generated probability forecasts |
| Weiyang Yan | Sample Empirical Likelihood Inference for Non-probability Samples |
Session 1B (Rm: Columbia B)
Chair: Qianyu Chen
| Dynamic Models | |
|---|---|
| Omar Essakine | Multivariate compound risk model based on tree-structured Poisson Markov random fields: Asymptotic results and risk sharing |
| Brandon Tam | Pure-Exchange Equilibria with Trade Aversion |
| Lilian Hu | Efficient and accurate simulation of the stochastic-alpha-beta-rho model |
12:15 p.m. | Lunch
Lunch will be served in the Main Hall.
1:30 p.m. | Parallel Session 2
Session 2A (Rm: Columbia A)
Chair: Yixin Ding
| Neural Networks | |
|---|---|
| Daniel Zhang | Tune Small, Train Big: BNN-μP for Bayesian Neural Network Language Models |
| Yixing Zhao | Neural Regression Trees for Conditional Distribution Learning under Censoring |
| Zihao Zhen | Neural Network Surrogates for Gradient-Based Bayesian Inference in Electrochemical Parameter Estimation |
| Taj Jones-McCormick | Towards Smarter LLM Powered Prediction Systems |
Session 2B (Rm: Columbia B)
Chair: Maria Andraos
| Modeling, Pricing, and Reserving | |
|---|---|
| Yan Yu | Variable Annuity Pricing with Limited Attention |
| Kathleen Miao | Consensus Copulas: Kullback-Leibler Barycentres for Model Aggregation |
| Tianxing Yan | Event-structured dependence in frequency-severity modelling of storm losses |
| Zhiyang Cheng | Neural Bayes Estimation for GLM-Based Loss Reserving |
3:10 p.m. | Coffee Break
Coffee will be served in the Main Hall.
3:30 p.m. | Research Workshop
Chair: Jay Sivathayalan
Presented by Frank Xingyu Chen & Xiao Ling Liu, Pacific Life Re
Topic: Mortality modelling and its application in industry.
4:00 p.m. | Parallel Session 3
Session 3A (Rm: Columbia A)
Chair: Xamy Lopez Cortes
| Meta Analysis | |
|---|---|
| Diribsa Tsegaye Bedada | Correcting Verification Bias in Individual Participant Data Meta-Analysis of Diagnostic Test Accuracy Studies: A Joint Likelihood Approach |
| Carmen Alicia Liang Zhen | Visualizing Joint Ranking Distributions in Network Meta-Analysis with Hammock Plots |
| Olana Angesa Dabi | A Meta-Analytic Model for Diagnostic Tests with Repeated Measurements |
Session 3B (Rm: Columbia B)
Chair: Qianyu Chen
| Risk Measures | |
|---|---|
| Muqiao Huang | Lambda Expected Shortfall |
| Jingcheng Yu | E-backtesting Expected Shortfall without Value-at-Risk forecast |
| Chenxi Xia | Rating Criteria, Risk Measures, and Risk Consistency |
5:20 p.m. | Reception (Rm: Main Hall)
The reception will take place in the Main Hall of FED Hall from 5:20 p.m. - 8:00 p.m.
9:00 a.m. | Breakfast (Rm: Main Hall)
9:30 a.m. | Keynote (Rm: Main Hall)
Chair: Laura Bumbulis
Dr. Erica Moodie, McGill University
Leveraging partial adherence to improve estimation of individualized treatments
Precision medicine is focused on optimizing sequential decision-making; in the statistical literature, this is formalized through dynamic treatment regimes. A flexible approach to estimation relies on inverse probability weighting. However, standard inverse weighted estimators can exhibit high variability. A key reason for this inefficiency arises because of the estimator’s requirement of strict alignment between observed and candidate treatment strategies which results in discarding substantial information from individuals with only slight deviations from the candidate strategies. In this talk, I will introduce two novel approaches that relax the strict alignment, resulting in estimators that are asymptotically consistent with the target estimand and offer good finite sample properties.
10:30 a.m. | Coffee Break
Coffee will be served in the Main Hall.
11:00 a.m. | Parallel Session 4
Session 4A (Rm: Columbia A)
Chair: Diribsa Bedada
| Machine Learning | |
|---|---|
| Yi Chang | Semiparametric Positive-Unlabeled Learning for Optimal Combination of Biomarkers |
| Nafisat Ibrahim | Identifying Factors Influencing TB Knowledge Among Nigerian Women Using Supervised Learning |
| Elijah French | Bayesian Learning in Graphon Games |
Session 4B (Rm: Columbia B)
Chair: Maria Andraos
| Portfolio Selection & Optimal Reinsurance | |
|---|---|
| Yuqi Jing | Adaptive Combination of Portfolio Rules |
| Yunqi Zhang | Optimal Reinsurance to Minimize the Drawdown Risk: A Probabilistic Approximation and Policy Iteration (PAPI) Approach |
| Brian Ceco | Active Portfolio Management in Concentrated Equity Markets |
12:15 p.m. | Lunch
Lunch will be served in the Main Hall.
1:30 p.m. | Parallel Session 5
Session 5A (Rm: Columbia A)
Chair: Yixin Ding
| E-Values | |
|---|---|
| Guneet Dhillon | E-Scores for (In)Correctness Assessment of Generative Model Outputs |
| Liulei Sun | Admissibility and Complete Classes for False Discovery Rate Control with E-values |
| Chenxin Lyu | E-Values |
Session 5B (Rm: Columbia B)
Chair: Madison Cranstone
| Causal Inference and Precision Medicine | |
|---|---|
| Gujie Fu | Causal Inference via Balanced Representation Learning |
| Xamy Lopez Cortes | Distributional Effects for Compliers under Instrumental Variables with Treatment Misclassification |
2:45 p.m. Coffee break
Coffee will be served in the Main Hall.