Online Consistency of the Nearest Neighbor Rule

Authors: Geelon So, Sanjoy Dasgupta

NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We prove online consistency for all measurable functions in doubling metric spaces under the mild assumption that the instances are generated by a process that is uniformly absolutely continuous with respect to a finite, upper doubling measure.
Researcher Affiliation Academia Sanjoy Dasgupta Department of Computer Science UC San Diego La Jolla, CA 92023 dasgupta@ucsd.edu; Geelon So Department of Computer Science UC San Diego La Jolla, CA 92023 geelon@ucsd.edu
Pseudocode Yes Algorithm 1 The 1-nearest neighbor rule 1: for n = 1, 2, . . . do 2: Receive the instance Xn 3: Predict with a nearest neighbor label η( Xn) 4: Observe and memorize the ground-truth label η(Xn) 5: end for
Open Source Code No The paper is theoretical and does not mention providing any open-source code.
Open Datasets No This is a theoretical work and does not involve experiments or datasets. The NeurIPS Paper Checklist also indicates 'NA' for questions related to experimental results and data.
Dataset Splits No This is a theoretical work and does not involve experiments or datasets, therefore no training/test/validation splits are discussed.
Hardware Specification No This is a theoretical work and does not involve experiments, thus no hardware specifications are mentioned.
Software Dependencies No This is a theoretical work and does not involve experiments, thus no software dependencies are listed.
Experiment Setup No This is a theoretical work and does not involve experiments, thus no experimental setup details like hyperparameters or training settings are provided.