Francesco Ragusa, Rosario Leonardi, Michele Mazzamuto
et al.
Understanding human behavior from complementary egocentric (ego) and exocentric (exo) points of view enables the development of systems that can support workers in industrial environments and enhance their safety. However, progress in this area is hindered by the lack of datasets capturing both views in realistic industrial scenarios. To address this gap, we propose ENIGMA-360, a new ego-exo dataset acquired in a real industrial scenario. The dataset is composed of 180 egocentric and 180 exocentric procedural videos temporally synchronized offering complementary information of the same scene. The 360 videos have been labeled with temporal and spatial annotations, enabling the study of different aspects of human behavior in industrial domain. We provide baseline experiments for 3 foundational tasks for human behavior understanding: 1) Temporal Action Segmentation, 2) Keystep Recognition and 3) Egocentric Human-Object Interaction Detection, showing the limits of state-of-the-art approaches on this challenging scenario. These results highlight the need for new models capable of robust ego-exo understanding in real-world environments. We publicly release the dataset and its annotations at https://fpv-iplab.github.io/ENIGMA-360/.
Industrial Cyber-Physical Systems (CPS) are sensitive infrastructure from both safety and economics perspectives, making their reliability critically important. Machine Learning (ML), specifically deep learning, is increasingly integrated in industrial CPS, but the inherent complexity of ML models results in non-transparent operation. Rigorous evaluation is needed to prevent models from exhibiting unexpected behaviour on future, unseen data. Explainable AI (XAI) can be used to uncover model reasoning, allowing a more extensive analysis of behaviour. We apply XAI to improve predictive performance of ML models intended for an industrial CPS use-case. We analyse the effects of components from time-series data decomposition on model predictions using SHAP values. Through this method, we observe evidence on the lack of sufficient contextual information during model training. By increasing the window size of data instances, informed by the XAI findings for this use-case, we are able to improve model performance.
AbstractPublic education systems and the incidence of child labor have historically been intertwined with both ultimately impacting labor market outcomes and the experience of work. This paper analyzes a suite of interrelated policies in the United States (some enacted, some proposed) that will have the ultimate effect of increasing the presence of minors in the workforce. We explore the impacts of this ultimate result for both industrial-organizational (I-O) research and practice, focusing on (a) increased underemployment and (b) increased workplace accidents, injuries and hazards in the workplace as clear points for necessary research and practice. Further, we highlight the need for I-O psychologists to become more adept at conducting research and practice with minors.
Tim Schreiter, Andrey Rudenko, Jens V. Rüppel
et al.
Successful adoption of industrial robots will strongly depend on their ability to safely and efficiently operate in human environments, engage in natural communication, understand their users, and express intentions intuitively while avoiding unnecessary distractions. To achieve this advanced level of Human-Robot Interaction (HRI), robots need to acquire and incorporate knowledge of their users' tasks and environment and adopt multimodal communication approaches with expressive cues that combine speech, movement, gazes, and other modalities. This paper presents several methods to design, enhance, and evaluate expressive HRI systems for non-humanoid industrial robots. We present the concept of a small anthropomorphic robot communicating as a proxy for its non-humanoid host, such as a forklift. We developed a multimodal and LLM-enhanced communication framework for this robot and evaluated it in several lab experiments, using gaze tracking and motion capture to quantify how users perceive the robot and measure the task progress.
Over the years, online scams have grown dramatically, with nearly 50% of global consumers encountering scam attempts each week. These scams cause not only significant financial losses to individuals and businesses, but also lasting psychological trauma, largely due to scammers' strategic employment of psychological techniques (PTs) to manipulate victims. Meanwhile, scammers continually evolve their tactics by leveraging advances in Large Language Models (LLMs) to generate diverse scam variants that easily bypass existing defenses. To address this pressing problem, we introduce PsyScam, a benchmark designed to systematically capture the PTs employed in real-world scam reports, and investigate how LLMs can be utilized to generate variants of scams based on the PTs and the contexts provided by these scams. Specifically, we collect a wide range of scam reports and ground its annotations of employed PTs in well-established cognitive and psychological theories. We further demonstrate LLMs' capabilities in generating through two downstream tasks: scam completion, and scam augmentation. Experimental results show that PsyScam presents significant challenges to existing models in both detecting and generating scam content based on the PTs used by real-world scammers. Our code and dataset are available.
Psychological counseling is a highly personalized and dynamic process that requires therapists to continuously monitor emotional changes, document session insights, and maintain therapeutic continuity. In this paper, we introduce PsyCounAssist, a comprehensive AI-powered counseling assistant system specifically designed to augment psychological counseling practices. PsyCounAssist integrates multimodal emotion recognition combining speech and photoplethysmography (PPG) signals for accurate real-time affective analysis, automated structured session reporting using large language models (LLMs), and personalized AI-generated follow-up support. Deployed on Android-based tablet devices, the system demonstrates practical applicability and flexibility in real-world counseling scenarios. Experimental evaluation confirms the reliability of PPG-based emotional classification and highlights the system's potential for non-intrusive, privacy-aware emotional support. PsyCounAssist represents a novel approach to ethically and effectively integrating AI into psychological counseling workflows.
Although extended reality(XR)-using technologies have started to be discussed in the industrial setting, it is becoming important to understand how to implement them ethically and privacy-preservingly. In our paper, we summarise our experience of developing XR implementations for the off-highway machinery domain by pointing to the main challenges we identified during the work. We believe that our findings can be a starting point for further discussion and future research regarding privacy and ethical challenges in industrial applications of XR.
Theofanis P. Raptis, Andrea Passarella, Marco Conti
Energy efficiency and reliability are vital design requirements of recent industrial networking solutions. Increased energy consumption, poor data access rates and unpredictable end-to-end data access latencies are catastrophic when transferring high volumes of critical industrial data in strict temporal deadlines. These requirements might become impossible to meet later on, due to node failures, or excessive degradation of the performance of wireless links. In this paper, we focus on maintaining the network functionality required by the industrial, best effort, low-latency applications after such events, by sacrificing latency guarantees to improve energy consumption and reliability. We avoid continuously recomputing the network configuration centrally, by designing an energy efficient, local and distributed path reconfiguration method. Specifically, given the operational parameters required by the applications, our method locally reconfigures the data distribution paths, when a network node fails. Additionally, our method also regulates the return to an operational state of nodes that have been offline in the past. We compare the performance of our method through simulations to the performance of other state of the art protocols and we demonstrate performance gains in terms of energy consumption, data delivery success rate, and in some cases, end-to-end data access latency. We conclude by providing some emerging key insights which can lead to further performance improvements.
Code completion, a crucial practice in industrial settings, helps developers improve programming efficiency by automatically suggesting code snippets during development. With the emergence of Large Code Models (LCMs), this field has witnessed significant advancements. Due to the natural differences between open-source and industrial codebases, such as coding patterns and unique internal dependencies, it is a common practice for developers to conduct domain adaptation when adopting LCMs in industry. There exist multiple adaptation approaches, among which retrieval-augmented generation (RAG) and fine-tuning are the two most popular paradigms. However, no prior research has explored the trade-off of the two approaches in industrial scenarios. To mitigate the gap, we comprehensively compare the two paradigms including Retrieval-Augmented Generation (RAG) and Fine-tuning (FT), for industrial code completion in this paper. In collaboration with Tencent's WXG department, we collect over 160,000 internal C++ files as our codebase. We then compare the two types of adaptation approaches from three dimensions that are concerned by industrial practitioners, including effectiveness, efficiency, and parameter sensitivity, using six LCMs. Our findings reveal that RAG, when implemented with appropriate embedding models that map code snippets into dense vector representations, can achieve higher accuracy than fine-tuning alone. Specifically, BM25 presents superior retrieval effectiveness and efficiency among studied RAG methods. Moreover, RAG and fine-tuning are orthogonal and their combination leads to further improvement. We also observe that RAG demonstrates better scalability than FT, showing more sustained performance gains with larger scales of codebase.
Katharina Leifker, Mathias Diebig, Ute Poethke
et al.
Introduction/Main Objectives: This study considered neural processes of transformational leadership based on quantitative electroencephalography (qEEG). Background Problems: This research aims at providing biomarkers for effective (i.e., transformational) leadership. Novelty: We considered transformational leadership on a detailed level, namely its individual-focused and group-focused sub-dimensions, to analyze the underlying brain processes. As for the individual-focused sub-dimensions of transformational leadership, we utilized innovation and performance orientation, while for the group-focused sub-dimensions, we choose vision and team spirit. Research Methods: Fifty-two dyads, consisting of (a) student pairs and (b) supervisor-subordinate dyads, participated in a simulated role-play that was intended to be a performance review while the electrical activity of the brain was recorded. Finding/Results: Results show that the group-focused sub-dimensions of transformational leadership could be positively linked to right frontal lobe coherence and negatively linked to left frontal lobe coherence. Results showed no relation between the individual-focused sub-dimensions and frontal lobe coherence. Conclusion: The results allow for a deeper understanding of the neural processes of transformational leadership and its individual-focused and group-focused sub-dimensions, respectively.
Purpose: Responsible leadership has gained significant attention in contemporary management scholarship. Despite significant advances made in the burgeoning corpus of knowledge, a paucity of studies focusses on developing economies with a dearth of studies underscoring specifically the public sector.
Design/methodology/approach: A qualitative research methodology was implemented and operationalised by means of a policy analysis. Ascribed to leadership activities within the public sector being linked to policy directives, a policy document analysis was deemed an appropriate approach. Content analysis was utilised as a data analytical method which consists of a systematic evaluation of the content of written or verbal communication for the purpose of identifying themes.
Findings/results: Findings are captured in a conceptual framework, which infers that leadership in the context of the public sector is consistent with the tenets of responsible leadership. Themes that emerged from the analysis include leadership antecedents, leadership roles, motivational drivers, and responsibility towards internal and external stakeholders. Leadership antecedents include psychological characteristics, aptitude and conative orientation. Normative, relational and operational roles were evidently stimulated by normative and intrapsychic motivation.
Practical implications: The contribution relates to how responsible leadership in terms of theoretical tenets and practical applicability might support policy makers and public sector leaders to improve the implementation of the National Development Plan (NDP) as a societal meta-framework for sustainable development in a multi-stakeholder context.
Originality/value: A dearth of research focuses on responsible leadership in the context of a developmental state especially theorising within the public sector indicative of the originality of the research.
The purpose of this research study was to study the influence of key psychological factors on emergence of Agile team autonomy that leads to Agile project success in software organizations.
While deep neural networks have achieved remarkable performance, data augmentation has emerged as a crucial strategy to mitigate overfitting and enhance network performance. These techniques hold particular significance in industrial manufacturing contexts. Recently, image mixing-based methods have been introduced, exhibiting improved performance on public benchmark datasets. However, their application to industrial tasks remains challenging. The manufacturing environment generates massive amounts of unlabeled data on a daily basis, with only a few instances of abnormal data occurrences. This leads to severe data imbalance. Thus, creating well-balanced datasets is not straightforward due to the high costs associated with labeling. Nonetheless, this is a crucial step for enhancing productivity. For this reason, we introduce ContextMix, a method tailored for industrial applications and benchmark datasets. ContextMix generates novel data by resizing entire images and integrating them into other images within the batch. This approach enables our method to learn discriminative features based on varying sizes from resized images and train informative secondary features for object recognition using occluded images. With the minimal additional computation cost of image resizing, ContextMix enhances performance compared to existing augmentation techniques. We evaluate its effectiveness across classification, detection, and segmentation tasks using various network architectures on public benchmark datasets. Our proposed method demonstrates improved results across a range of robustness tasks. Its efficacy in real industrial environments is particularly noteworthy, as demonstrated using the passive component dataset.
Cross-Domain Recommendation (CDR) have received widespread attention due to their ability to utilize rich information across domains. However, most existing CDR methods assume an ideal static condition that is not practical in industrial recommendation systems (RS). Therefore, simply applying existing CDR methods in the industrial RS environment may lead to low effectiveness and efficiency. To fill this gap, we propose DIIT, an end-to-end Domain-Invariant Information Transfer method for industrial cross-domain recommendation. Specifically, We first simulate the industrial RS environment that maintains respective models in multiple domains, each of them is trained in the incremental mode. Then, for improving the effectiveness, we design two extractors to fully extract domain-invariant information from the latest source domain models at the domain level and the representation level respectively. Finally, for improving the efficiency, we design a migrator to transfer the extracted information to the latest target domain model, which only need the target domain model for inference. Experiments conducted on one production dataset and two public datasets verify the effectiveness and efficiency of DIIT.
Background: With eHealth technology interventions, users' personal health data can be easily shared among different stakeholders. Users should decide with whom they want to share their data. As support, most eHealth technology has data sharing options functionalities. However, there is little research on how to design these visually. In this paper, we took two possible data sharing options designs - data and party perspective – for an existing eHealth technology intervention, and we explored them. Objective: The aim was to find which of the two designs is the best in terms of trust, privacy concerns, ease of use, and information control. Additionally, to investigate how these factors influence each other with also the goal of giving practical advice on designing for privacy. Method: We conducted a between-subjects online design experiment (N = 123). After having visualised one of the two data sharing options designs, participants filled in an online questionnaire. To analyse the data, t-test analyses, correlation analyses, and backward regression analyses were conducted. Results: Information control scored higher in the data perspective condition (t (97) = 2.25, p = .03). From the different regression analyses, we found that trust and ease of use play a role in all sharing-related factors. Conclusions: We concluded that the design of data-sharing options in eHealth technology affects the experience of the user, mostly for trust and ease of use. In the end, we provided several actionable design advices on how to design for privacy.
Artificial intelligence agents (chatbots), which are programs to communicate with users primarily in customer service contexts, are an alternative interaction channel supporting businesses in the digital environment and vital components in customer service. The present empirical paper, which is aimed at identifying and discussing the factors motivating nonusers to adopt the specific technology in mobile contexts, proposes a comprehensive conceptual model, which combines the UTAUT 2 behavioral theory with variables of mobile service quality contexts, such as information quality, privacy concerns, interface, and equipment, as well as trust and mobility factors. Data analysis, based on the partial least squares structural equation modeling (PLS-SEM) statistical method, revealed that performance expectancy, facilitating factors, hedonic motivation, mobility, trust, and service quality positively affect nonusers’ behavioral intention to adopt chatbots. In addition, equipment, interface, and trust have a significant impact on users’ trust in the context of mobile chatbots. Personal data privacy issues also have a negative effect on trust, in contrast to effort expectancy, which positively affects performance expectancy. As mobile service quality factors have not been investigated before in the context of chatbots, the findings of the present research are expected to provide useful insights both to academia and the business industry.
Zilong Yin,1,* Wanying Guo,2,* Yue Wang3 1School of Teacher Education, Jiangsu University, Zhenjiang, People’s Republic of China; 2Overseas Education College, Jiangsu University, Zhenjiang, People’s Republic of China; 3College of Education, Chungnam National University, Daejeon, Republic of Korea*These authors contributed equally to this workCorrespondence: Zilong Yin, School of Teacher Education, Jiangsu University, No. 301 Xuefu Road, Zhenjiang, Jiangsu Province, 212013, People’s Republic of China, Email yinzilong1@hotmail.comPurpose: The relationship between psychological distress (PD) and academic engagement (AE) in adolescents is widely recognized. However, there is a lack of research on the mechanisms behind this correlation, especially in COVID-19 pandemic policy liberalization. This study proposed a moderated mediation model to determine the mediating role of social-emotional competence (SEC) between PD and AE and the moderating role of teachers’ caring behavior (TCB).Patients and methods: The moderated mediation model was tested by PROCESS macro developed based on SPSS with Chinese adolescents aged 13– 18 years (n = 8625; male 50.5%) as survey respondents. Multiple scales, including the 6-item Kessler PD Scale, SEC Scale, TCB Scale, and AE Scale, were used in this test.Results: Adolescent PD negatively predicted AE. In addition, SEC partially mediated the relationship between adolescent PD and AE. TCB moderated the direct and indirect relationships between PD and AE, which mitigated the negative effect of PD on AE and enhanced the positive effect of SEC on AE.Conclusion: Adolescent SEC and TCB are essential in mediating PD and AE. This study provided instructions for schools and teachers to intervene in adolescents’ PD and academic problems, such as improving SEC and engaging in TCB in an adequate and timely manner. In addition, these findings provided vital references for adolescent mental health interventions in future large-scale public health events.Keywords: psychological distress, academic engagement, social-emotional competence, teachers’ caring behavior, adolescents, COVID-19 policy liberalization
Ekram Goyal, Alka Puria, Suprakash Chaudhary
et al.
Introduction: Chronic Kidney Disease (CKD) is a debilitating illness that impairs an individual's physical and social functioning and ultimately affects the quality of life (QOL). Aim: To determine the impact of psychiatric comorbidity on QOL and activities of daily living in individuals suffering from CKD undergoing hemodialysis. Materials and Methods: Fifty subjects suffering from CKD undergoing hemodialysis were consequently enrolled in the department of medicine. Institutional ethics committee permission was obtained before the start of the study. The interview was conducted only after the hemodialysis procedure. Consent was obtained and socio-demographic details were noted in the socio-demographic proforma. Schedule for Clinical Assessment in Neuropsychiatry (SCAN) was used to assess psychiatric comorbidity. QOL was assessed on Short Form-36 (SF-36). Katz Index of Independence of Activities of Daily Living scale (KADL) was used to assess the activities of living of such individuals. Results: 46% (n = 23) of subjects suffering from CKD had psychiatric co-morbidity. The most common psychiatric disorder observed was moderate depressive disorder (14%, n = 7) followed by adjustment disorder (12%, n = 6). QOL on SF-36 in all eight domains was low compared to the general population. There was a high statistically significant negative impact of psychiatric comorbidity on QOL on eight domains of SF-36 and activities of daily living. (P = 0.001). Conclusion: Psychiatric co-morbidity is common in CKD patients on hemodialysis. Quality of life is lower compared to the general population. There is a negative impact of psychiatric co-morbidity on quality of life and activities of daily living.
Muhammad Abbas, Ali Hamayouni, Mahshid Helali Moghadam
et al.
Processing and reviewing nightly test execution failure logs for large industrial systems is a tedious activity. Furthermore, multiple failures might share one root/common cause during test execution sessions, and the review might therefore require redundant efforts. This paper presents the LogGrouper approach for automated grouping of failure logs to aid root/common cause analysis and for enabling the processing of each log group as a batch. LogGrouper uses state-of-art natural language processing and clustering approaches to achieve meaningful log grouping. The approach is evaluated in an industrial setting in both a qualitative and quantitative manner. Results show that LogGrouper produces good quality groupings in terms of our two evaluation metrics (Silhouette Coefficient and Calinski-Harabasz Index) for clustering quality. The qualitative evaluation shows that experts perceive the groups as useful, and the groups are seen as an initial pointer for root cause analysis and failure assignment.