Navegando por Data de Publicação, começando com "2025-09-08"

Filtrar resultados por ano ou mês

Agora exibindo 1 - 3 de 3

As atividades significativas e o cotidiano das pessoas em sofrimento psíquico pelo uso de álcool e outras drogas: perspectivas em Terapia Ocupacional
(Universidade Federal de São Carlos, 2025-09-08) Carleto, Daniel Gustavo de Sousa; Lussi, Isabela Aparecida de Oliveira; https://lattes.cnpq.br/8121264125922144; https://lattes.cnpq.br/9571012795056991; Marcolino, Tais Quevedo; Fernandes, Amanda Dourado Souza Akahosi; Pereira, Andrea Ruzzi; Leão, Adriana; https://lattes.cnpq.br/7197838515800679; https://lattes.cnpq.br/7901666626822625; https://lattes.cnpq.br/2764871991850142; https://lattes.cnpq.br/5298816927945151
Drug use has been a part of humanity since ancient times, present in different cultures with different motivations. Harmful use can lead to fragility and the rupture of social bonds, isolation, and vulnerability in various aspects of people's lives. In this context, Occupational Therapy works to promote social reintegration and a meaningful daily life. Through activities that are meaningful to the individual, the aim is to rescue abandoned experiences and create possibilities, aiding in their recovery and social reintegration. This doctoral thesis aimed to understand the meaningful activities and daily lives of people suffering from psychological distress due to alcohol and other drug use from their own perspectives. The research, which used a qualitative approach, involved 10 people suffering from psychological distress due to alcohol and other drug use who attended a Psychosocial Care Center for Alcohol and Drugs in a municipality in the Triângulo Mineiro region. The theoretical framework used was the theory of everyday life proposed by Agnes Heller. Data were collected over a five-month period, specifically from February to July 2023, through a sociodemographic questionnaire and an adaptation of the Photovoice method. This consisted of asking participants to take photos of meaningful activities they perform, would like to perform, represent their daily lives, and would like to be part of their daily lives. The data collected through the sociodemographic questionnaire were analyzed descriptively. The data collected through the adapted Photovoice method and the interview guide were analyzed using thematic content analysis, from which three categories of analysis emerged: What is meaningful to do: human activities; A daily life of meaningful activities; Using drugs is a meaningful activity (yes)! The results showed that participants attributed different meanings to meaningful activities and daily life, including the use of psychoactive substances itself. Daily life was understood as a daily routine composed of various activities, in addition to care/assistance activities, interpersonal relationships, and work activities. The data revealed that drug use influences the performance of meaningful activities and the daily lives of participants, but that it can also be considered a meaningful activity in people's lives. The research findings demonstrated the relevance of the Alcohol and Drug Psychosocial Care Center and the development of feelings of belonging in the contexts in which people with psychological distress due to alcohol and other drug use live for the performance of meaningful activities and the redefinition of their daily lives. Based on the theoretical framework adopted, it was possible to analyze the influence of psychological distress and drug use on the daily lives of participants. The study demonstrates the importance of considering the perspectives of people with psychological distress due to alcohol and other drug use in valuing their meaningful activities and redefining their daily lives, enabling qualification for this area of Occupational Therapy.
Sistema de reconhecimento de fala disártrica usando aprendizagem autossupervisionada
(Universidade Federal de São Carlos, 2025-09-08) Gracelli, Ricardo Alexandre; Almeida Junior, Jurandy Gomes de; https://lattes.cnpq.br/4495269939725770; https://orcid.org/0000-0002-4998-6996; https://lattes.cnpq.br/6620281314745744; https://orcid.org/0009-0007-7457-0603; Comin, Cesar Henrique; Santos, Thiago Oliveira dos; https://lattes.cnpq.br/9563440403120931; https://lattes.cnpq.br/5117339495064254
This study aims to develop and evaluate Automatic Speech Recognition (ASR) systems tailored to the needs of individuals with dysarthric speech — a condition that compromises communication clarity and limits the use of voice-based assistive technologies. One of the main challenges in dysarthric speech recognition lies in the scarcity of labeled data, and to address this issue, two complementary and interdependent approaches were examined. The first investigated pathology-oriented data augmentation techniques applied to two Transformer-based processing pipelines: FW1 and FW2. Signal perturbation methods — additive noise, time-stretching, and the proposed Spectral Oclution (SO) — were applied individually and in combination to recordings from the UA-Speech corpus. Exploratory analysis of Word Recognition Accuracy (WRA), Word Error Rate (WER), and Character Error Rate (CER) curves showed that combining noise and time-stretching consistently reduced errors in speakers with moderate intelligibility; SO provided additional gains in specific cases; and the union of all three perturbations benefited severe cases, although it could degrade nearly typical voices due to spectral overheating. The second approach, built upon the findings of the first, employed supervised pre-training on typical speech from the LJSpeech corpus. These baseline models were then subjected to a self-supervised contrastive cycle on the typical partition of UA-Speech, where the same transformations from Phase 1 (noise, time-stretching, and SO) were reused to generate positive pairs for contrastive training. Thus, augmentation strategies were not only validated in isolation but also served as the foundation for the contrastive stage. We evaluated two methods: Simple Framework for Contrastive Learning of Visual Representations (SimCLR) and Swapping Assignments between Views (SwAV). The refined weights were then transferred for fine-tuning on dysarthric datasets. Compared to the baseline trained solely on LJSpeech, both contrastive methods enhanced performance: SimCLR showed higher sensitivity to severe speakers, while SwAV maintained stable performance across all intelligibility levels, further reducing CER and WER and increasing WRA. In some severe cases, WER was reduced by more than 20 percentage points. In summary, the integration of targeted data augmentation and contrastive pre-training resulted in ASR models more robust to the articulatory variability of dysarthria, supporting the inclusion of dysarthric speakers in voice-based communication systems.
O papel da hesitação na interação humano-máquina: uma análise no atendimento via chat
(Universidade Federal de São Carlos, 2025-09-08) Santos, Valeria Vieira; Stassi-Sé, Joceli Catarina; https://lattes.cnpq.br/1314068157460143; https://lattes.cnpq.br/4618767941163669; https://orcid.org/0009-0006-0023-6736; Fontes, Michel Gustavo; Vale, Oto Araújo; https://lattes.cnpq.br/1816530656343898; https://lattes.cnpq.br/2277403284693571
This research aims to describe and analyze hesitation in interactions carried out through customer service chats that use Artificial Intelligence (AI) systems, based on the Interactive Textual Perspective (Jubran, 2006). The investigation focuses on identifying and describing the linguistic markers of hesitation in interactions mediated by three types of interlocutors: human agents, users, and a chatbot, as well as verifying how these markers manifest and impact the perception of fluency, comprehension, and humanization in dialogue. The study is grounded in the understanding of AI as a technical field that also seeks to develop agents capable of simulating human behavior (Russell; Norvig, 2016; Nilsson, 1998), with particular focus on automated conversational systems (Jurafsky; Martin, 2023), such as chatbots, which constitute hybrid environments between orality and writing. In this context, hesitation is defined as an interactive-discursive phenomenon with a textual function, capable of signaling doubt, organizing discourse, and seeking cooperation, as discussed by Marcuschi (1999), Marcuschi (2000). The analyzed data are drawn from a corpus of real customer service interactions conducted by a nationally recognized company that authorized the use of the data. The methodology involves the qualitative and quantitative categorization of hesitation markers according to linguistic criteria, including frequency of hesitations, type of hesitation, communicative context, impact on interactions, user response, human vs. AI comparison, conversation length, pragmatics of hesitation, cognitive aspect of hesitation, and interactional aspect of hesitation. The analysis shows that hesitation, when more frequent among human interlocutors—mainly in functional words and hesitation markers—is associated with interactions that generate effects of empathy and cooperation from users. In contrast, the absence or superficial simulation of this phenomenon in AI turns compromises the continuity and naturalness of the interaction. By situating hesitation as an important element in the construction and fluency of conversations mediated by humans and robots, this research explores and connects the subject to the field of technology and may contribute to studies on the humanization of AI in automated customer service, reinforcing the importance of integrating linguistic, pragmatic, and technological aspects in the design of conversational systems that are more sensitive to natural language.