Optimal Clustering with Noisy Queries via Multi-Armed Bandit

Authors: Jinghui Xia, Zengfeng Huang

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

Reproducibility Variable Result LLM Response
Research Type Theoretical In this work, we obtain the first matching upper and lower bounds for a wide range of parameters. In particular, a new polynomial time algorithm with O( n(k+log n)/δ2 + poly(k, 1/δ , log n)) queries is proposed. Moreover, we prove a new lower bound of Ω( n log n/δ2 )
Researcher Affiliation Academia 1School of Data Science, Fudan University, Shanghai, China 2Shanghai Key Lab of Intelligent Information Processing. Correspondence to: Zengfeng Huang <huangzf@fudan.edu.cn>.
Pseudocode Yes Algorithm 1 Sub Cluster(V, k, δ, b): sub-clusters recovery, Algorithm 2 True Cluster Id(u, X1, ..., Xk, δ, α), Algorithm 3 Cluster Verify(v, B), Algorithm 4 Bal Noisy Clustering(V, k, δ, b), Algorithm 5 Gap SBM(T, h, δ), Algorithm 6 Noisy Clustering(V, k, δ), Algorithm 7 A s: algorithm for BAIF.
Open Source Code No The paper does not contain any statement about making its source code available or provide a link to a code repository.
Open Datasets No The paper describes a theoretical framework for clustering with noisy queries and does not involve real-world datasets or a training phase.
Dataset Splits No The paper is theoretical and does not describe experimental setups with data splits (train/validation/test).
Hardware Specification No The paper describes theoretical algorithms and proofs, and does not mention any hardware used for running experiments.
Software Dependencies No The paper is theoretical and does not provide details on specific software dependencies or their version numbers.
Experiment Setup No The paper is theoretical and does not describe an experimental setup with hyperparameters or training configurations.