Iskanje po katalogu digitalne knjižnice

Opis: Because of the prevalence of depression, its often-chronic course, relapse and associated disability, early detection and non-intrusive monitoring is a crucial tool for timely diagnosis and treatment, remission of depression and prevention of relapse. In this way, its impact on quality of life and well-being can be limited. Current attempts to use artificial intelligence for the early classification of depression are mostly data-driven and thus non-transparent and lack effective means to deal with uncertainties. Therefore, in this paper, we propose an end-to-end framework for extracting observable depression cues from diary recordings. Furthermore, we also explore its feasibility for automatic detection of depression symptoms using observable behavioural cues. The proposed end-to-end framework for extracting depression was used to evaluate 28 video recordings from the Symptom Media dataset and 27 recordings from the DAIC-WOZ dataset. We compared the presence of the extracted features between recordings of individuals with and without a depressive disorder. We identified several cues consistent with previous studies in terms of their differentiation between individuals with and without depressive disorder across both datasets among language (i.e., use of negatively valanced words, use of first-person singular pronouns, some features of language complexity, explicit mentions of treatment for depression), speech (i.e., monotonous speech, voiced speech and pauses, speaking rate, low articulation rate), and facial cues (i.e., rotational energy of head movements). The nature/context of the discourse, the impact of other disorders and physical/psychological stress, and the quality and resolution of the recordings all play an important role in matching the digital features to the relevant background. In this way, the work presented in this paper provides a novel approach to extracting a wide range of cues relevant to the classification of depression and opens up new opportunities for further research.
Ključne besede: digital biomarkers of depression, facial cues, speech cues, language cues, deep learning, end-to-end pipeline, artificial intelligence
Objavljeno v DKUM: 17.01.2025; Ogledov: 0; Prenosov: 0
Celotno besedilo (2,34 MB)

Opis: Exposure to false information is becoming a common occurrence in our daily lives. New developments in artificial intelligence are now used to produce increasingly sophisticated multimedia false content, such as deepfakes, making false information even more challenging to detect and combat. This creates expansive opportunities to mislead individuals into believing fabricated claims and negatively influence their attitudes and behavior. Therefore, a better understanding of how individuals perceive such content and the variables related to the perceived trustworthiness of deepfakes is needed. In the present study, we developed and validated the Perceived Deepfake Trustworthiness Questionnaire (PDTQ) in English, Italian, and Slovene. This was done in three phases. First, we developed the initial pool of items by reviewing previous studies, generating items via interviews and surveys, and employing artificial intelligence. Second, we shortened and adapted the questionnaire according to experts’ evaluation of content validity and translated the questionnaire into Italian and Slovene. Lastly, we evaluated the psychometric characteristics via a cross-sectional study in three languages (N ¼ 733). The exploratory factor analyses suggested a two-factor solution, with the first factor measuring the perceived trustworthiness of the content and the second measuring the perceived trustworthiness of the presentation. This factorial structure was replicated in confirmatory factor analyses. Moreover, our analyses provided support for PDTQ’s reliability, measurement invariance across all three languages, and its construct and incremental validity. As such, the PDTQ is a reliable, measurement invariant, and valid tool for comprehensive exploration of individuals’ perception of deepfake videos.
Ključne besede: deepfakes, misinformation, perception, questionnaire validation, trust
Objavljeno v DKUM: 03.09.2024; Ogledov: 30; Prenosov: 8
Celotno besedilo (2,12 MB)
Gradivo ima več datotek! Več...

Opis: Healthcare systems around the world are currently witnessing various challenges, including population aging and workforce shortages. As a result, the existing, overworked staff are struggling to meet the ever-increasing demands and provide the desired quality of care. One of the promising technological solutions that could complement the human workforce and alleviate some of their workload, are socially assistive humanoid robots. However, despite their potential, the implementation of socially assistive humanoid robots is often challenging due to low acceptance among key stakeholders, namely, patients and healthcare professionals. Hence, the present study first investigated the extent to which these stakeholders accept the use of socially assistive humanoid robots in nursing and care routine, and second, explored the characteristics that contribute to higher/lower acceptance within these groups, with a particular emphasis on demographic variables, technology expectations, ethical acceptability, and negative attitudes. In study 1, conducted on a sample of 490 healthcare professionals, the results of structural equation modeling showed that acceptance is driven primarily by aspects of ethical acceptability, although education and technology expectations also exert an indirect effect. In study 2, conducted on a sample of 371 patients, expectations regarding capabilities and attitudes towards the social influence of robots emerged as important predictors of acceptance. Moreover, although acceptance rates differed between tasks, both studies show a relatively high acceptance of socially assistive humanoid robots. Despite certain limitations, the study findings provide essential knowledge that enhances our understanding of stakeholders’ perceptions and acceptance of socially assistive humanoid robots in hospital environments, and may guide their deployment.
Ključne besede: socially assistive humanoid robots, attitudes, expectations, patients, healthcare professionals
Objavljeno v DKUM: 12.06.2024; Ogledov: 166; Prenosov: 14
Celotno besedilo (1,74 MB)
Gradivo ima več datotek! Več...

Ključne besede: corpora and language resources, speech corpus, multimodal corpus, pragmatics, conversational intelligence
Objavljeno v DKUM: 10.04.2024; Ogledov: 285; Prenosov: 10
Celotno besedilo (2,08 MB)
Gradivo ima več datotek! Več...

Opis: Small and medium enterprises are significantly hampered by cyber-threats as they have inherently limited skills and financial capacities to anticipate, prevent, and handle security incidents. The EU-funded PALANTIR project aims at facilitating the outsourcing of the security supervision to external providers to relieve SMEs/MEs from this burden. However, good practices for the operation of SME/ME assets involve avoiding their exposure to external parties, which requires a tightly defined and timely enforced security policy when resources span across the cloud continuum and need interactions. This paper proposes an innovative architecture extending Network Function Virtualisation to externalise and automate threat mitigation and remediation in cloud, edge, and on-premises environments. Our contributions include an ontology for the decision-making process, a Fault-and-Breach-Management-based remediation policy model, a framework conducting remediation actions, and a set of deployment models adapted to the constraints of cloud, edge, and on-premises environment(s). Finally, we also detail an implementation prototype of the framework serving as evaluation material.
Ključne besede: Security-as-a-Service, security orchestration, policy-driven management, virtual network functions, finite state machines, constraints programming
Objavljeno v DKUM: 06.02.2024; Ogledov: 334; Prenosov: 16
Celotno besedilo (963,42 KB)
Gradivo ima več datotek! Več...