RT Journal Article
T1 Early stopping by correlating online indicators in neural networks
A1 Vilares Ferro, Manuel
A1 Doval Mosquera, Yerai
A1 Ribadas Pena, Francisco Jose
A1 Darriba Bilbao, Victor Manuel
K1 1203.04 Inteligencia Artificial
AB In order to minimize the generalization error in neural networks, a novel technique to identifyoverfitting phenomena when training the learner is formally introduced. This enables support of areliable and trustworthy early stopping condition, thus improving the predictive power of that typeof modeling. Our proposal exploits the correlation over time in a collection of online indicators,namely characteristic functions for indicating if a set of hypotheses are met, associated with a range ofindependent stopping conditions built from a canary judgment to evaluate the presence of overfitting.That way, we provide a formal basis for decision making in terms of interrupting the learning process.As opposed to previous approaches focused on a single criterion, we take advantage of subsidiaritiesbetween independent assessments, thus seeking both a wider operating range and greater diagnosticreliability. With a view to illustrating the effectiveness of the halting condition described, we chooseto work in the sphere of natural language processing, an operational continuum increasingly based onmachine learning. As a case study, we focus on parser generation, one of the most demanding andcomplex tasks in the domain. The selection of cross-validation as a canary function enables an actualcomparison with the most representative early stopping conditions based on overfitting identification,pointing to a promising start toward an optimal bias and variance control.
PB Neural Networks
SN 08936080
YR 2023
FD 2023-02
LK http://hdl.handle.net/11093/4277
UL http://hdl.handle.net/11093/4277
LA eng
NO Neural Networks, 159, 109-124 (2023)
NO Financiado para publicación en acceso aberto: Universidade de Vigo/CISUG
NO info:eu-repo/grantAgreement/AEI/Plan Estatal de Investigación Científica y Técnica y de Innovación 2013-2016/TIN2017-85160-C2-2-R/ES/AVANCES EN NUEVOS SISTEMAS DE EXTRACCION DE RESPUESTAS CON ANALISIS SEMANTICO Y APRENDIZAJE PROFUNDO
NO info:eu-repo/grantAgreement/AEI/Plan Estatal de Investigación Científica y Técnica y de Innovación 2017-2020/PID2020-113230RB-C22/ES/SEQUENCE LABELING MULTITASK MODELS FOR LINGUISTICALLY ENRICHED NER: SEMANTICS AND DOMAIN ADAPTATION (SCANNER-UVIGO)
NO Agencia Estatal de Investigación | Ref. TIN2017-85160-C2-2-R
DS Investigo
RD 07-dic-2023