Delayed Feedback in Kernel Bandits

Authors: Sattar Vakili, Danyal Ahmed, Alberto Bernacchia, Ciara Pike-Burke

ICML 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We also validate our theoretical results with simulations.
Researcher Affiliation Collaboration 1Media Tek Research 2Imperial College London.
Pseudocode Yes Algorithm 1 Batch Pure Exploration with Delays (BPE-Delay)
Open Source Code Yes The code for these experiments is provided in a Git Hub repository.2 2https://github.com/svakili89/delayed kernel bandit
Open Datasets No The paper states: 'These functions are generated by fitting a kernel based model to points randomly generated from a multivariate Gaussian.' This indicates generated data for simulation, not a publicly available dataset with concrete access information for training.
Dataset Splits No The paper discusses simulation experiments using generated data rather than standard public datasets with predefined train/validation/test splits. It does not specify exact split percentages or sample counts for validation.
Hardware Specification No The paper does not specify any particular hardware used for running the experiments (e.g., specific GPU or CPU models, memory details, or cloud instance types).
Software Dependencies No The paper does not list specific version numbers for software dependencies (e.g., Python, PyTorch, or other libraries).
Experiment Setup Yes We use a SE kernel with a length scale parameter l = 0.8 for f1 and l = 1.0 for f2 in order to generate these objective functions. The learner can then choose from |X| = 2500 points over a uniform 50 50 grid. The sampling noise is zero mean Gaussian with standard deviation σ = 0.02. The stochastic delay in the feedback is generated from a Poisson distribution with parameter λ. The calculation of ψt for BPE-Delay uses ξ = 9 and b = 1 given in Assumption 4.2.