On Kernelized Multi-Armed Bandits with Constraints

Authors: Xingyu Zhou, Bo Ji

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate the superior performance of our proposed algorithms via numerical experiments based on both synthetic and real-world datasets.
Researcher Affiliation Academia Xingyu Zhou Electrical and Computer Engineering Wayne State University Detroit, MI, USA xingyu.zhou@wayne.edu Bo Ji Computer Science Virginia Tech Blacksburg, VA, USA boji@vt.edu
Pseudocode Yes Algorithm 1 CKB Algorithm
Open Source Code Yes 3. If you ran experiments... (a) Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [Yes]
Open Datasets Yes Light-Tailed Real-World Data. We use the light sensor data collected in the CMU Intelligent Workplace in Nov 2005, which is available online as Matlab structure3 and contains locations of 41 sensors, 601 train samples and 192 test samples. ... 3http://www.cs.cmu.edu/~guestrin/Class/10708-F08/projects/
Dataset Splits No The paper specifies '601 train samples and 192 test samples' for the light sensor data, but does not explicitly mention a validation split or provide information for all three splits (train/validation/test).
Hardware Specification No Experiments were rather run on a personal computer.
Software Dependencies No The paper does not provide specific software dependencies with version numbers.
Experiment Setup Yes The objective function f( ) = Pp i=1 aik( , xi) is generated by uniformly sampling ai 2 [ 1, 1] and support points xi 2 X with p = 100. With the same manner, we generate the constraint function g. The kernel is kse with parameter l = 0.2. Other parameters include B, R and γt are set similar as in the unconstrained case (e.g., [5]). ... The constraint is given by g( ) = f( )+h with h = B/2.