Information-theoretic Limits of Online Classification with Noisy Labels

Authors: Changlong Wu, Ananth Grama, Wojciech Szpankowski

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

Reproducibility Variable Result LLM Response
Research Type Theoretical Our main contributions in this paper establish the fundamental limits of minimax risk in (3) by providing nearly matching lower and upper bounds across a wide range of hypothesis classes H and noisy kernels K. This is a pure theory paper.
Researcher Affiliation Academia Changlong Wu Ananth Grama Wojciech Szpankowski CSo I, Purdue University wuchangl@hawaii.edu, {ayg,szpan}@purdue.edu
Pseudocode Yes Algorithm 1: Predictor via Pairwise Hypothesis Testing and Algorithm 2: Exponential Weighted Average (EWA) estimator
Open Source Code No This is a pure theory paper. No statement about releasing open-source code for the methodology described in the paper is found.
Open Datasets No This is a pure theory paper and does not involve the use of a dataset for training or evaluation.
Dataset Splits No This is a pure theory paper and does not specify training, validation, or test dataset splits.
Hardware Specification No This is a pure theory paper and does not specify any hardware used for running experiments.
Software Dependencies No This is a pure theory paper and does not specify any software dependencies with version numbers.
Experiment Setup No This is a pure theory paper and does not include details about an experimental setup, such as hyperparameters or training configurations.