Supporting Complaints Investigation for Nursing and Midwifery Regulatory Agencies

Lertvittayakumjorn, Piyawat; Petej, Ivan; Gao, Yang; Krishnamurthy, Yamuna; Gaag, Anna van der; Jago, Robert; Stathis, Kostas

doi:10.18653/v1/2021.acl-demo.10

Cited by 4 publications

(3 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Existirán casos en los que será necesario realizar esta tarea de forma manual [21]. Frecuentemente los modelos presentan problemas de sesgo, esto se debe a muchos factores, sin embargo, uno importante a tener en cuenta es la correcta distribución de los datos tal es el caso de la investigación hecha por [22], ellos pre procesaron los textos de tal forma que el género estuviese correctamente balanceado.…”

Section: Figura 3 Coocurrencia De Palabras De Acuerdo a Complaint Cla...unclassified

Clasificación de quejas de los clientes empleando procesamiento de lenguaje natural: revisión sistemática de la literatura

Flores Poma

2023

Rev peru comput sist

View full text Add to dashboard Cite

Un cliente insatisfecho por algún producto y/o servicio se encuentra motivado a expresar una queja. Clasificar las quejas de forma manual es un proceso que representa elevados costos en recursos humanos y materiales. La Inteligencia Artificial (IA) permite el uso de diversos algoritmos para realizar tareas que pueden simular la inteligencia humana, una rama de esta es el Procesamiento de Lenguaje Natural (NLP por sus siglas en inglés), tiene por objetivo que las máquinas tengan la capacidad de entender lenguaje humano, permitiendo, por ejemplo clasificar y categorizar datos de forma automática. Este artículo proporciona una revisión sistemática de la literatura abordando desafíos en la clasificación de textos de queja, tales como la falta de balance de clases, la presencia de datos sin etiquetar y la interpretación de los resultados de los modelos. Se exploran técnicas de pre procesamiento, como la tokenización, la remoción de stopwords y la lematización, que influyen en el rendimiento de los modelos. Adicionalmente, se discuten las métricas de rendimiento como precision, recall y F1-score. Se muestran las tendencias actuales y futuras líneas de investigación. Para tal fin se analizaron 24 artículos publicados entre 2018 y 2023 extraídas de las bases de datos de Web of Science y Scopus.

show abstract

Section: Figura 3 Coocurrencia De Palabras De Acuerdo a Complaint Cla...unclassified

Clasificación de quejas de los clientes empleando procesamiento de lenguaje natural: revisión sistemática de la literatura

Flores Poma

2023

Rev peru comput sist

View full text Add to dashboard Cite

show abstract

“…Humans have been using explanations in AI for many purposes such as to support human decision making [35,39], to increase human trust in the AI [27,65], to verify and improve the AI [10,40], and to learn new knowledge from the AI [34,44]. Explanations may also be required for an AI-assisted system to comply with recent regulations including the General Data Protection Regulation (GDPR) [21].…”

Section: Introductionmentioning

confidence: 99%

Argumentative explanations for pattern-based text classifiers

Lertvittayakumjorn

Toni

2023

AAC

Self Cite

View full text Add to dashboard Cite

Recent works in Explainable AI mostly address the transparency issue of black-box models or create explanations for any kind of models (i.e., they are model-agnostic), while leaving explanations of interpretable models largely underexplored. In this paper, we fill this gap by focusing on explanations for a specific interpretable model, namely pattern-based logistic regression (PLR) for binary text classification. We do so because, albeit interpretable, PLR is challenging when it comes to explanations. In particular, we found that a standard way to extract explanations from this model does not consider relations among the features, making the explanations hardly plausible to humans. Hence, we propose AXPLR, a novel explanation method using (forms of) computational argumentation to generate explanations (for outputs computed by PLR) which unearth model agreements and disagreements among the features. Specifically, we use computational argumentation as follows: we see features (patterns) in PLR as arguments in a form of quantified bipolar argumentation frameworks (QBAFs) and extract attacks and supports between arguments based on specificity of the arguments; we understand logistic regression as a gradual semantics for these QBAFs, used to determine the arguments’ dialectic strength; and we study standard properties of gradual semantics for QBAFs in the context of our argumentative re-interpretation of PLR, sanctioning its suitability for explanatory purposes. We then show how to extract intuitive explanations (for outputs computed by PLR) from the constructed QBAFs. Finally, we conduct an empirical evaluation and two experiments in the context of human-AI collaboration to demonstrate the advantages of our resulting AXPLR method.

show abstract

“…Humans have been using explanations in AI for many purposes such as to support human decision making [1,2], to increase human trust in the AI [3,4], to verify and improve the AI [5,6], and to learn new knowledge from the AI [7,8]. Explanations may also be required for an AI-assisted system to comply with recent regulations including the General Data Protection Regulation (GDPR) [9].…”

Section: Introductionmentioning

confidence: 99%

Argumentative Explanations for Pattern-Based Text Classifiers

Lertvittayakumjorn¹,

Toni²

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Recent works in Explainable AI mostly address the transparency issue of black-box models or create explanations for any kind of models (i.e., they are model-agnostic), while leaving explanations of interpretable models largely underexplored. In this paper, we fill this gap by focusing on explanations for a specific interpretable model, namely pattern-based logistic regression (PLR) for binary text classification. We do so because, albeit interpretable, PLR is challenging when it comes to explanations. In particular, we found that a standard way to extract explanations from this model does not consider relations among the features, making the explanations hardly plausible to humans. Hence, we propose AXPLR, a novel explanation method using (forms of) computational argumentation to generate explanations (for outputs computed by PLR) which unearth model agreements and disagreements among the features. Specifically, we use computational argumentation as follows: we see features (patterns) in PLR as arguments in a form of quantified bipolar argumentation frameworks (QBAFs) and extract attacks and supports between arguments based on specificity of the arguments; we understand logistic regression as a gradual semantics for these QBAFs, used to determine the arguments' dialectic strength; and we study standard properties of gradual semantics for QBAFs in the context of our argumentative re-interpretation of PLR, sanctioning its suitability for explanatory purposes. We then show how to extract intuitive explanations (for outputs computed by PLR) from the constructed QBAFs. Finally, we conduct an empirical evaluation and two experiments in the context of human-AI collaboration to demonstrate the advantages of our resulting AXPLR method.

show abstract

Supporting Complaints Investigation for Nursing and Midwifery Regulatory Agencies

Cited by 4 publications

References 41 publications

Clasificación de quejas de los clientes empleando procesamiento de lenguaje natural: revisión sistemática de la literatura

Clasificación de quejas de los clientes empleando procesamiento de lenguaje natural: revisión sistemática de la literatura

Argumentative explanations for pattern-based text classifiers

Argumentative Explanations for Pattern-Based Text Classifiers

Contact Info

Product

Resources

About