2023
Aguilera, Aurora Cobo; Olmos, Pablo M; Artés-Rodríguez, Antonio; Pérez-Cruz, Fernando
Regularizing transformers with deep probabilistic layers Artículo de revista
En: Neural Networks, 2023, ISSN: 0893-6080.
Resumen | Enlaces | BibTeX | Etiquetas: Deep learning, Missing data, Natural language processing, Regularization, Transformers, Variational auto-encoder
@article{AGUILERA2023,
title = {Regularizing transformers with deep probabilistic layers},
author = {Aurora Cobo Aguilera and Pablo M Olmos and Antonio Art\'{e}s-Rodr\'{i}guez and Fernando P\'{e}rez-Cruz},
url = {https://www.sciencedirect.com/science/article/pii/S0893608023000448},
doi = {https://doi.org/10.1016/j.neunet.2023.01.032},
issn = {0893-6080},
year = {2023},
date = {2023-01-01},
urldate = {2023-01-01},
journal = {Neural Networks},
abstract = {Language models (LM) have grown non-stop in the last decade, from sequence-to-sequence architectures to attention-based Transformers. However, regularization is not deeply studied in those structures. In this work, we use a Gaussian Mixture Variational Autoencoder (GMVAE) as a regularizer layer. We study its advantages regarding the depth where it is placed and prove its effectiveness in several scenarios. Experimental result demonstrates that the inclusion of deep generative models within Transformer-based architectures such as BERT, RoBERTa, or XLM-R can bring more versatile models, able to generalize better and achieve improved imputation score in tasks such as SST-2 and TREC or even impute missing/noisy words with richer text.},
keywords = {Deep learning, Missing data, Natural language processing, Regularization, Transformers, Variational auto-encoder},
pubstate = {published},
tppubtype = {article}
}
Sedano-Capdevila, Alba; Toledo-Acosta, Mauricio; Barrigon, María Luisa; Morales-González, Eliseo; Torres-Moreno, David; Martínez-Zaldivar, Bolívar; Hermosillo-Valadez, Jorge; Baca-García, Enrique; Aroca, Fuensanta; Artes-Rodriguez, Antonio; Baca-García, Enrique; Berrouiguet, Sofian; Billot, Romain; Carballo-Belloso, Juan Jose; Courtet, Philippe; Gomez, David Delgado; Lopez-Castroman, Jorge; Rodriguez, Mercedes Perez; Aznar-Carbone, Julia; Cegla, Fanny; Gutiérrez-Recacha, Pedro; Izaguirre-Gamir, Leire; Herrera-Sanchez, Javier; Borja, Marta Migoya; Palomar-Ciria, Nora; Martínez, Adela Sánchez-Escribano; Vasquez, Manuel; Vallejo-Oñate, Silvia; Vera-Varela, Constanza; Amodeo-Escribano, Susana; Arrua, Elsa; Bautista, Olga; Barrigón, Maria Luisa; Carmona, Rodrigo; Caro-Cañizares, Irene; Carollo-Vivian, Sonia; Chamorro, Jaime; González-Granado, Marta; Iza, Miren; Jiménez-Giménez, Mónica; López-Gómez, Ana; Mata-Iturralde, Laura; Miguelez, Carolina; Muñoz-Lorenzo, Laura; Navarro-Jiménez, Rocío; Ovejero, Santiago; Palacios, María Luz; Pérez-Fominaya, Margarita; Peñuelas-Calvo, Inmaculada; Pérez-Colmenero, Sonia; Rico-Romano, Ana; Rodriguez-Jover, Alba; SánchezAlonso, Sergio; Sevilla-Vicente, Juncal; Vigil-López, Carolina; Villoria-Borrego, Lucía; Martin-Calvo, Marisa; Alcón-Durán, Ana; Stasio, Ezequiel Di; García-Vega, Juan Manuel; Martín-Calvo, Pedro; Ortega, Ana José; Segura-Valverde, Marta; Bañón-González, Sara María; Crespo-Llanos, Edurne; Codesal-Julián, Rosana; Frade-Ciudad, Ainara; Merino, Elena Hernando; Álvarez-García, Raquel; Coll-Font, Jose Marcos; Antonio, Pablo Portillo-de; Puras-Rico, Pablo; Sedano-Capdevila, Alba; Serrano-Marugán, Leticia
Text mining methods for the characterisation of suicidal thoughts and behaviour Artículo de revista
En: Psychiatry Research, vol. 322, pp. 115090, 2023, ISSN: 0165-1781.
Resumen | Enlaces | BibTeX | Etiquetas: Machine learning, Mobile health, Natural language processing, Suicidal ideation, Suicide, Suicide attempt
@article{SEDANOCAPDEVILA2023115090,
title = {Text mining methods for the characterisation of suicidal thoughts and behaviour},
author = {Alba Sedano-Capdevila and Mauricio Toledo-Acosta and Mar\'{i}a Luisa Barrigon and Eliseo Morales-Gonz\'{a}lez and David Torres-Moreno and Bol\'{i}var Mart\'{i}nez-Zaldivar and Jorge Hermosillo-Valadez and Enrique Baca-Garc\'{i}a and Fuensanta Aroca and Antonio Artes-Rodriguez and Enrique Baca-Garc\'{i}a and Sofian Berrouiguet and Romain Billot and Juan Jose Carballo-Belloso and Philippe Courtet and David Delgado Gomez and Jorge Lopez-Castroman and Mercedes Perez Rodriguez and Julia Aznar-Carbone and Fanny Cegla and Pedro Guti\'{e}rrez-Recacha and Leire Izaguirre-Gamir and Javier Herrera-Sanchez and Marta Migoya Borja and Nora Palomar-Ciria and Adela S\'{a}nchez-Escribano Mart\'{i}nez and Manuel Vasquez and Silvia Vallejo-O\~{n}ate and Constanza Vera-Varela and Susana Amodeo-Escribano and Elsa Arrua and Olga Bautista and Maria Luisa Barrig\'{o}n and Rodrigo Carmona and Irene Caro-Ca\~{n}izares and Sonia Carollo-Vivian and Jaime Chamorro and Marta Gonz\'{a}lez-Granado and Miren Iza and M\'{o}nica Jim\'{e}nez-Gim\'{e}nez and Ana L\'{o}pez-G\'{o}mez and Laura Mata-Iturralde and Carolina Miguelez and Laura Mu\~{n}oz-Lorenzo and Roc\'{i}o Navarro-Jim\'{e}nez and Santiago Ovejero and Mar\'{i}a Luz Palacios and Margarita P\'{e}rez-Fominaya and Inmaculada Pe\~{n}uelas-Calvo and Sonia P\'{e}rez-Colmenero and Ana Rico-Romano and Alba Rodriguez-Jover and Sergio S\'{a}nchezAlonso and Juncal Sevilla-Vicente and Carolina Vigil-L\'{o}pez and Luc\'{i}a Villoria-Borrego and Marisa Martin-Calvo and Ana Alc\'{o}n-Dur\'{a}n and Ezequiel Di Stasio and Juan Manuel Garc\'{i}a-Vega and Pedro Mart\'{i}n-Calvo and Ana Jos\'{e} Ortega and Marta Segura-Valverde and Sara Mar\'{i}a Ba\~{n}\'{o}n-Gonz\'{a}lez and Edurne Crespo-Llanos and Rosana Codesal-Juli\'{a}n and Ainara Frade-Ciudad and Elena Hernando Merino and Raquel \'{A}lvarez-Garc\'{i}a and Jose Marcos Coll-Font and Pablo Portillo-de Antonio and Pablo Puras-Rico and Alba Sedano-Capdevila and Leticia Serrano-Marug\'{a}n},
url = {https://www.sciencedirect.com/science/article/pii/S0165178123000434},
doi = {https://doi.org/10.1016/j.psychres.2023.115090},
issn = {0165-1781},
year = {2023},
date = {2023-01-01},
journal = {Psychiatry Research},
volume = {322},
pages = {115090},
abstract = {Traditional research methods have shown low predictive value for suicidal risk assessments and limitations to be applied in clinical practice. The authors sought to evaluate natural language processing as a new tool for assessing self-injurious thoughts and behaviors and emotions related. We used MEmind project to assess 2838 psychiatric outpatients. Anonymous unstructured responses to the open-ended question “how are you feeling today?” were collected according to their emotional state. Natural language processing was used to process the patients' writings. The texts were automatically represented (corpus) and analyzed to determine their emotional content and degree of suicidal risk. Authors compared the patients' texts with a question used to assess lack of desire to live, as a suicidal risk assessment tool. Corpus consists of 5,489 short free-text documents containing 12,256 tokenized or unique words. The natural language processing showed an ROC-AUC score of 0.9638 when compared with the responses to lack of a desire to live question. Natural language processing shows encouraging results for classifying subjects according to their desire not to live as a measure of suicidal risk using patients’ free texts. It is also easily applicable to clinical practice and facilitates real-time communication with patients, allowing better intervention strategies to be designed.},
keywords = {Machine learning, Mobile health, Natural language processing, Suicidal ideation, Suicide, Suicide attempt},
pubstate = {published},
tppubtype = {article}
}