Regret Bounds for Gaussian-Process Optimization in Large Domains

Authors: Manuel Wuethrich, Bernhard Schölkopf, Andreas Krause

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical It is not the purpose of the present paper to propose a novel algorithm with better performance (we modify the EI and UCB strategies merely to facilitate the proof), but rather to gain an understanding of how GP-optimization performs in the aforementioned setting, as a function of the number of evaluations and the domain size.
Researcher Affiliation Academia Manuel W uthrich MPI for Intelligent Systems T ubingen, Germany Bernhard Sch olkopf MPI for Intelligent Systems T ubingen, Germany Andreas Krause ETH Zurich Zurich, Switzerland
Pseudocode No The paper defines EI2 (Definition 2) and UCB2 (Definition 3) with mathematical formulas, but it does not include a clearly labeled pseudocode or algorithm block.
Open Source Code No The paper does not provide an unambiguous statement or a direct link to the source code for the methodology described in the paper. It references third-party tools like GPy and Scikit-learn, but not its own code.
Open Datasets No The paper focuses on theoretical analysis of Gaussian Processes and regret bounds. It defines a "problem instance" by parameters (N, T, µ, Σ) or (A, T, µ, k), not by using or providing access to specific empirical datasets.
Dataset Splits No The paper does not mention training, validation, or test dataset splits, as it is a theoretical paper that does not conduct empirical experiments on specific datasets.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., GPU/CPU models, memory) used to conduct any computations or generate the results presented.
Software Dependencies No The paper mentions GPy and Scikit-learn in the references, implying they are related tools or frameworks in the field, but it does not explicitly list them as software dependencies with specific version numbers required to replicate the paper's analytical work or any implicit computations.
Experiment Setup No The paper is theoretical and does not describe an empirical experimental setup with hyperparameters, training configurations, or system-level settings.