Опубликована: Ноя. 8, 2024
Язык: Английский
Опубликована: Ноя. 8, 2024
Язык: Английский
IEEE Access, Год журнала: 2024, Номер 12, С. 75735 - 75760
Опубликована: Янв. 1, 2024
Machine learning (ML) has become a popular technique for various automation tasks in the era of Industry 4.0, such as analysis and synthesis visual data images videos, natural language speech, financial data, biomedical applications. However, ML-based techniques are facing difficulties like decision-making, thus incorporating user expertise into system might be advantageous. The goal adding human domain with is to provide more accurate prediction models. Human-in-the-loop (HITL) systems that integrate ML algorithms becoming common industries. there number methodological, technical, ethical development application HITL systems. This paper aims explore methodologies, challenges, opportunities associated implementations.We also discuss issues must resolved effective, including quality, bias, engagement. Besides, we explored several approaches can utilized enhance performance systems, active (AL), iterative ML, reinforcement learning, well current state art systems.We selectively highlighted advantages their potential increase decision-making process accountability transparency by utilizing experience improve capability. will very useful researchers, practitioners, policymakers.
Язык: Английский
Процитировано
14Microchemical Journal, Год журнала: 2024, Номер 200, С. 110250 - 110250
Опубликована: Фев. 27, 2024
Язык: Английский
Процитировано
7Advanced Engineering Informatics, Год журнала: 2024, Номер 61, С. 102504 - 102504
Опубликована: Март 26, 2024
Язык: Английский
Процитировано
4Advances in computational intelligence and robotics book series, Год журнала: 2024, Номер unknown, С. 354 - 383
Опубликована: Авг. 15, 2024
The chapter “AI Safety and Security” presents a comprehensive multi-dimensional exploration, addressing the critical aspects of safety security in context large language models. begins by identifying risks threats posed LLMs, delving into vulnerabilities such as bias, misinformation, unintended AI interactions, impacts like privacy concerns. Building on these identified risks, it then explores strategies methodologies for ensuring safety, focusing principles robustness, transparency, accountability discussing challenges implementing measures. It concludes with an insight long-term research, highlighting ongoing efforts future directions to sustain system amidst rapid technological advancements encouraging collaborative approach among various stakeholders. By integrating perspectives from computer science, ethics, law, social sciences, provides insightful analysis current security.
Язык: Английский
Процитировано
3Journal of Systems and Software, Год журнала: 2025, Номер 223, С. 112373 - 112373
Опубликована: Фев. 13, 2025
Язык: Английский
Процитировано
0IFLA Journal, Год журнала: 2025, Номер unknown
Опубликована: Фев. 13, 2025
Generative artificial intelligence tools are becoming ubiquitous in applications across personal, professional and educational contexts. Similar to the rise of social media technologies, this means they an embedded part people's lives, individuals using these for a variety benign purposes. This article examines how existing information literacy understandings will not work literacy, provides example searching, demonstrating its shortcomings. Present approaches may fall short answer required navigate new tools, begs question what comes next. The current scope technology necessitates multidisciplinary approach solving ‘what do with intelligence’ arguably most impactfully requires one acknowledge that has worked no longer suffice.
Язык: Английский
Процитировано
0Computer Communications, Год журнала: 2025, Номер unknown, С. 108144 - 108144
Опубликована: Март 1, 2025
Язык: Английский
Процитировано
0PeerJ Computer Science, Год журнала: 2025, Номер 11, С. e2795 - e2795
Опубликована: Апрель 15, 2025
This study presents an augmented hybrid approach for improving the diagnosis of malignant skin lesions by combining convolutional neural network (CNN) predictions with selective human interventions based on prediction confidence. The algorithm retains high-confidence CNN while replacing low-confidence outputs expert assessments to enhance diagnostic accuracy. A model utilizing EfficientNetB3 backbone is trained datasets from ISIC-2019 and ISIC-2020 SIIM-ISIC melanoma classification challenges evaluated a 150-image test set. model’s are compared against 69 experienced medical professionals. Performance assessed using receiver operating characteristic (ROC) curves area under curve (AUC) metrics, alongside analysis resource costs. baseline achieves AUC 0.822, slightly below performance experts. However, improves true positive rate 0.782 reduces false 0.182, delivering better minimal involvement. offers scalable, resource-efficient solution address variability in image analysis, effectively harnessing complementary strengths humans CNNs.
Язык: Английский
Процитировано
0Internet of Things, Год журнала: 2024, Номер 27, С. 101234 - 101234
Опубликована: Май 31, 2024
Язык: Английский
Процитировано
1International Journal of Human-Computer Interaction, Год журнала: 2024, Номер unknown, С. 1 - 18
Опубликована: Окт. 14, 2024
This study investigates effective feedback mechanisms to maintain human engagement in interactive machine learning (IML) systems, focusing on social media platforms. We developed "Loop," an IML system based human-in-the-loop (HITL) principles that recommends content while encouraging users report inaccuracies for model refinement. Loop implements three types of artificial intelligence (AI) user reports: (a) (ML)-centric, (b) personal-centric, and (c) community-centric feedback. In addition, we evaluated the relative effectiveness these under two different task criticality scenarios: high low. A with 30 participants was conducted evaluate through questionnaires interviews. Results showed preferred algorithmic improvements personal benefit over altruistic contributions community, especially low-criticality tasks. Furthermore, personal-centric had a significant impact satisfaction. Our findings provide insights into HITL-ML contributing design more engaging interfaces. discuss implications strategies proactive HITL-ML-based emphasizing importance tailored mechanisms.
Язык: Английский
Процитировано
1