The day in conversation for 2026-11-04
discussion
Transcript
The discussion: Tom: I’m still circling back to that Gaussian-efficient testing paper; the idea of betting directly on the mean of bounded data seems incredibly powerful, especially when we look at how it handles iid data where we usually rely on asymptotic normality. Jane: That efficiency claim is compelling for large samples, but I'm curious about the practical implementation details—does that strategy remain tractable when the underlying distribution isn't perfectly bounded, or are we constrained by that assumption to achieve the stated Gaussian efficiency? Lu: From a different angle, while we discuss testing efficiency, I wonder how this relates to our earlier work on rerandomization; is there an overlap in how those finite-sample gains for difference-in-means estimators might inform the bounds you’re setting for these new confidence intervals? Meng: I think the connection lies in robustness. If we look at the causal excursion effect estimation paper, which showed relative efficiencies up to three, it suggests that achieving high efficiency isn't just about getting a tight interval for one parameter;
More episodes
- Daily Summary for 2026-11-19
- 2310.07973-Statistical Performance Guarantee for Subgroup Identification with Generic Machine Learning
- 2408.09155-Learning Robust Treatment Rules for Censored Data
- Daily Summary for 2026-11-19
- 2603.23397-Kinetic Langevin Splitting Schemes for Constrained Sampling
- 2606.09906-An information-geometric framework for mapping maximum potential biodiversity
- 2603.27718-Induced replication and the assessment of models
- 1307.0898-Entropy of a Zipfian Distributed Lexicon
- 2406.19974-Generalizing self-normalized importance sampling with couplings
- 2412.10213-Collaborative Optimal Design of Controlled Experiments with Subject Covariates