Offensive-Language Detection on Multi-Semantic Fusion Based on Data Augmentation

The rapid identification of offensive language in social media is of great significance for preventing viral spread and reducing the spread of malicious information, such as cyberbullying and content related to self-harm. In existing research, the public datasets of offensive language are small; the label quality is uneven; and the performance of the pre-trained models is not satisfactory. To overcome these problems, we proposed a multi-semantic fusion model based on data augmentation (MSF). Data augmentation was carried out by back translation so that it reduced the impact of too-small datasets on performance. At the same time, we used a novel fusion mechanism that combines word-level semantic features and n-grams character features. The experimental results on the two datasets showed that the model proposed in this study can effectively extract the semantic information of offensive language and achieve state-of-the-art performance on both datasets.

Adaptation and Preliminary Validation of the Arabic Version of the Role Checklist Version 3 (RCv3)

Arab occupational therapists are in need either to develop their own outcome measures or to translate well-known outcome measures to their culture. The objective of this study was to translate, cross-culturally adapt, and validate preliminarily the Role Checklist Version 3 (RCv3) into Arabic language. The translation and cross-cultural adaptation process of the RCv3 into Arabic language was conducted in four main steps: (a) forward translation, (b) back translation, (c) expert panel review process, and (d) psychometric testing (Validity). The expert panel established the face validity and the content validity of the Arabic RCv3. Pilot testing was conducted, and the utility and content validity of the tool were confirmed. The final version of the Arabic RCv3 is now available for use in the Arab countries. However, further psychometric properties including reliability and responsiveness of this tool is warranted, thus supporting its wide use and proper application.

Spanish version of the Frontotemporal Dementia Knowledge Scale: adaptation and validation

ABSTRACT Background: Frontotemporal dementia (FTD) is a neurodegenerative disease and is one of the most common causes of dementia in people under 65. There is often a significant diagnostic delay, as FTD can be confused with other psychiatric conditions. A lack of knowledge regarding FTD by health professionals is one possible cause for this diagnostic confusion. Objectives: The aim of this study was to adapt and validate the Frontotemporal Dementia Knowledge Scale (FTDKS) in Spanish. Methods: A translation was done, following cross-cultural adaptation guidelines, which consisted of forward translation, blind back translation, and an analysis by a committee of experts. For the present study, 134 professionals from different health areas responded the Spanish version of the FTDKS. The statistical analysis was performed using R version 4.0.0 “Arbor day” and the Psych, sjPlot packages. Results: The Spanish version of the FTDKS had good reliability and internal consistency (Cronbach alpha 0.74.). The sample's mean score was 19.78 (range = 4-32, SD 6.3) out of a maximum of 36 points. Conclusions: The results obtained show that the Spanish version has good psychometric properties. The FTDKS is applicable in our environment and can be a useful tool to evaluate the knowledge of health professionals regarding frontotemporal dementia.

Human-Computer Interaction in Translation Activity: Fluency of Machine Translation

Digitalization is one of the key distinctive features of modern environment and social life. Nowadays more and more functions are transferred to the artificial mind. How effective is the replacement of human activity with computer activity? In the given article, this problem is solved by an example of integration of digital technologies into translation activities. It this paper, emphasis is placed on the quality of machine translation (MT) output of legal texts in the language pair English - Slovak. It studies a Criminal Code formulated in the Slovak language which was translated by a human translator into English and consequently via machine translation system Google Translate (GT) back into Slovak. The back-translation - translation of a translated text back into its original language - as a quality assessment tool to detect discrepancies, mistranslations and inevitable differences between the source text and the target text was used. The quality of MT output was evaluated according to Multidimensional Quality Metrics (MQM) standards with the focus on the dimension of Fluency. The multiple comparisons were applied to determine which issues (errors) in Fluency dimension differ from the others. A statistically significant difference is noticed between Agreement and other issues, as well as between Ambiguity and other issues. The errors in Agreement are related to the differences between the languages: English is considered mostly an analytic language, Slovak represents a synthetic language. The issues in the Ambiguity dimension correlate with the type of the text being examined, since legal texts are characterized by relatively complicated wording and numerous terms; moreover, accuracy and unambiguity need to be preserved. Generally, the MT output is able to provide users with basic information about the text. On the other hand, most of the segments need revision and/or correction; in such cases, human intervention and post-editing is necessary.

Evidences of validity of the Brief Resilience Scale for Brazil

Introduction: Resilience is the ability to adapt or recover after adverse situations.This study aimed to adapt and investigate evidence of the validity of the Brief Resilience Scale (BRS) for the Brazilian context. Method: The total sample comprised 1,937 people who participated in the study and the validity analysis was carried out with 1,480 people between 18 and 78 years of age. It was carried out using translation, back-translation, descriptive analysis, reliability analysis, confirmatory factor analysis, Item Response Theory, the ROC curve and BRS correlation with the Connor-Davidson Resilience Scale - 10-item version (CD-RISC-10), the Social Support Perception Scale, the Patient Health Questionnaire - 2-item version (PHQ-2), and sociodemographic data. Results: The results showed the adequacy of the Brazilian version of the BRS (BRS-BR), alpha of 0.80, omega = 0.81, one factor, good information capacity of the items (except for item 5) and correlation with the CD-RISC-10 (rho = 0.64), PHQ-2 (rho = - 0.38)and Social Support (rho = 0.14). There was also a correlation with sex (rho = 0.11), age (rho = 0.13), marital status (rho = 0.15) and schooling (rho = 0.15). The ROC curve shows a cutoff point at 10 points for low resilience and 21 points for high resilience. Conclusions: The investigation of the psychometric characteristics of the BRS-BR showed that the instrument can be useful for the evaluation of resilience in the Brazilian context.

Sentence Augmentation for Language Translation Using GPT-2

Data augmentation has recently become an important method for improving performance in deep learning. It is also a significant issue in machine translation, and various innovations such as back-translation and noising have been made. In particular, current state-of-the-art model architectures such as BERT-fused or efficient data generation using the GPT model provide good inspiration to improve the translation performance. In this study, we propose the generation of additional data for neural machine translation (NMT) using a sentence generator by GPT-2 that produces similar characteristics to the original. BERT-fused architecture and back-translation are employed for the translation architecture. In our experiments, the model produced BLEU scores of 27.50 for tatoebaEn-Ja, 30.14 for WMT14En-De, and 24.12 for WMT18En-Ch.

Us vs them: Tribalism, selective exposure & how non-partisan online news media engagement can be facilitated

<p>Codeswitching is the action of switching between languages to better impart information to a recipient. This thesis introduces a set of codeswitching translator tools as a method of disrupting the potentially damaging structures of tribal politics through the manipulation of ideologically specific language norms. We first investigate how tribalism and group identity impact our ability to participate in political discourse. Using this insight from a host of different research disciplines, we design an iterative testing environment for a variety of ‘codeswitching’ translators in order to see the impact of translations ranging in complexity from simple word and syntax substitution through to machine learning back-translation. Though back-translation was not found to be an effective technique, simple substitution methods provided a foundation of effectiveness and proof of concept among test participants, especially those that identified as politically aligned.</p>

Us vs them: Tribalism, selective exposure & how non-partisan online news media engagement can be facilitated

<p>Codeswitching is the action of switching between languages to better impart information to a recipient. This thesis introduces a set of codeswitching translator tools as a method of disrupting the potentially damaging structures of tribal politics through the manipulation of ideologically specific language norms. We first investigate how tribalism and group identity impact our ability to participate in political discourse. Using this insight from a host of different research disciplines, we design an iterative testing environment for a variety of ‘codeswitching’ translators in order to see the impact of translations ranging in complexity from simple word and syntax substitution through to machine learning back-translation. Though back-translation was not found to be an effective technique, simple substitution methods provided a foundation of effectiveness and proof of concept among test participants, especially those that identified as politically aligned.</p>

Linguistic Validation of Indonesian Version of the Audit of Diabetes-Dependent Quality of Life Questionnaire

Background: One of the impacts experienced by diabetes mellitus patients is a decrease in their quality of life. The Audit of Diabetes-Dependent Quality of Life (ADDQoL) is a widely used individualized diabetes-specific quality of life measure. However, there was no version available in the Indonesian language. Objective: This study is aimed to undertake linguistic validation, including a cultural adaptation of the ADDQoL questionnaire into the Indonesian language. Method: The original developer granted permission to use and modify the questionnaire. The international linguistic validation procedure developed by the Mapi Research Institute was used. There were six steps involved: forward translation, reconciliation, back translation, expert panel review by a psychologist and clinician, cognitive debriefing with diabetes patients, and proofreading. Result: Problems that arose during the linguistic validation process were resolved by finding conceptually equivalent alternatives and changing sentence structures to achieve equivalence in language, concept, and culture with the original version of the ADDQoL. The developer's team reviewed and discussed all actions taken. Cognitive debriefing interviews with five respondents showed that the ADDQoL questionnaire was simple to understand. Conclusion: The Indonesian version of the ADDQoL is linguistically and culturally validated. Further studies are needed to confirm the structure and reliability of the Indonesian ADDQoL.