On the ERM Principle With Networked Data

Authors: Yuanhong Wang, Yuyi Wang, Xingwu Liu, Juhua Pu

AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical in this paper we mainly prove new universal risk bounds for CLANET whose goal is to train a classifier with examples in a data graph G. formulate a non-convex optimization problem inspired by our new risk bounds and then we also design a new efficient algorithm to obtain an approximate optimal weighting vector and show that this algorithm is a fully polynomial-time approximation scheme for this non-convex program.
Researcher Affiliation Academia 1 State Key Laboratory of Software Development Environment, Beihang University, Beijing, China 2 Research Institute of Beihang University in Shenzhen, Shenzhen, China 3 Disco Group, ETH Zurich, Switzerland 4 Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China 5 University of Chinese Academy of Sciences, Beijing, China
Pseudocode Yes Algorithm 1 FPTAS for weighting vector optimization.
Open Source Code No The paper provides a link to an online appendix (https://arxiv.org/abs/1711.04297) for supplementary details, but does not explicitly state that source code for their method is released or available.
Open Datasets No The paper is theoretical and does not mention or provide access to any specific datasets for empirical training or evaluation.
Dataset Splits No The paper is theoretical and does not specify any training/validation/test dataset splits.
Hardware Specification No The paper is theoretical and does not specify any hardware used for experiments.
Software Dependencies No The paper is theoretical and does not list any specific software dependencies with version numbers.
Experiment Setup No The paper is theoretical and does not provide details about an experimental setup, hyperparameters, or training settings.