Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 10 of 10 for “"finite-sample analysis"”.
-
Finite-sample analysis for distributed and neural reinforcement learning
This dissertation develops a modular finite-sample theory for reinforcement learning under two realities of modern practice that challenge classical analysis: nonlinear function approximation and trajectory-based, temporally dependent data, including distributed and continuing settings. The central …
-
A Unified Lyapunov Framework for Finite-Sample Analysis of Reinforcement Learning Algorithms
… learning algorithms and establishing their finite-sample guarantees. Specifically, we aim at answering the following question: suppose we carry out some reinforcement learning algorithm with finite amount of samples (or with finite number of iterations), then what can we say about the …
-
Advances in Proximal Inference for Continuous Exposures, Estimation of Ill-Posed Regression, and Non-Inferiority Assessment in Active-Controlled Trials
… misspecifications. Chapter 3, motivated by the analysis of immune correlates of protection in COVID-19 vaccine trials, extends the proximal inference framework to identify and estimate the mean outcome under modified treatment policies, enabling causal analysis of continuous exposures in the …
-
Reinforcement Learning with Finite Time and Sample Guarantees
… demonstrations and human feedback, establishing finite-sample guarantees for reward estimation. I first present a single-loop regularized Q-learning algorithm for minimizing the projected Bellman error with finite time convergence guarantees in the case of linear function approximation. The …
-
Temporal abstraction and generalisation in reinforcement learning
… of state abstraction, temporal abstraction, and finite-data generalisation performance for reinforcement learning algorithms. To focus on data abstraction, we explore an imitation learning setting. We provide a novel algorithm for completely offline imitation learning, as well as an empirical …
-
Folded Variance Estimators for Stationary Time Series
This thesis is concerned with simulation output analysis. In particular, we are interested in estimating the variance parameter of a steady-state output process. The estimation of the variance parameter has immediate applications in problems involving (i) the precision of the sample mean as a point …
-
Towards Foundation Models for Text-Rich Multimodal Tabular Data
… guarantee on the upper bound on error using finite-sample analysis. Empirically, we demonstrated improvements on both the standard benchmark for multi-armed bandits and a newly proposed benchmark for realistic news recommendation tasks. On the question of input representations, we started …
-
Three Essays in Time Series Econometric Analysis
… Consistent (HAC) estimators with improved finite sample properties. This is achieved by exploiting existing transformations from the related Heteroskedasticity Consistent estimator literature. Monte Carlo simulation results provide strong evidence in favour of using these alternative …
-
On the information theory of clustering, registration, and blockchains
Progress in data science depends on the collection and storage of large volumes of reliable data, efficient and consistent inference based on this data, and trusting such computations made by untrusted peers. Information theory provides the means to analyze statistical inference algorithms, …
-
Latent Structure in Linear Prediction and Corpora Comparison
This work first studies the finite-sample properties of the risk of the minimum-norm interpolating predictor in high-dimensional regression models. If the effective rank of the covariance matrix of the p regression features is much larger than the sample size n, we show that the min-norm …