Earticle

Recently, neural architecture search (NAS) has received increasing attention as it can replace human experts in designing the architecture of neural networks for different tasks and has achieved remarkable results in many challenging tasks. In this study, a path-based computation neural architecture encoder (PCE) was proposed. Our PCE first encodes the computation of information on each path in a neural network, and then aggregates the encodings on all paths together through an attention mechanism, simulating the process of information computation along paths in a neural network and encoding the computation on the neural network instead of the structure of the graph, which is more consistent with the computational properties of neural networks. We performed an extensive comparison with eight encoding methods on two commonly used NAS search spaces (NAS-Bench-101 and NAS-Bench-201), which included a comparison of the predictive capabilities of performance predictors and search capabilities based on two search strategies (reinforcement learning-based and Bayesian optimization-based) when equipped with different encoders. Experimental evaluation shows that PCE is an efficient encoding method that effectively ranks and predicts neural architecture performance, thereby improving the search efficiency of neural architectures.

Purpose: Family caregivers of people living with dementia (PLWD) may experience mental and physical health issues. The purpose of this study was to examine factors associated with family caregiver outcomes and to describe their interrelationships. Methods: Grounded in the stress process model, path analysis was conducted using secondary data from 178 dyads of community-dwelling PLWD and family caregivers in Korea. Caregiver outcome was assessed through caregivers’ health status and stress due to behavioral and psychological symptoms of dementia (BPSD). Agitation and depression were specified as primary stressors, and PLWD factors were included as background factors associated with BPSD, with separate symptom models. Results: Agitation and depression showed the strongest associations with family caregiver stress related to BPSD. Agitation explained 45.1% of caregiver stress related to BPSD variance (β=.66, p<.001), while depression explained 41.5% (β=.63, p<.001). Agitation and depression were related to caregiver health status (β=.17, p=.006; β=.18, p=.017). In both agitation and depression models, female caregivers (β=.19, p=.008, in both models), older age (β=.21, p=.007, in both models), and lower economic status (β=.32, p<.001, in both models) were associated with poorer caregiver health status. Dementia severity and patient age showed indirect associations with caregiver stress related to BPSD through agitation and depression. Conclusion: Agitation and depression were associated with caregiver stress related to BPSD, while caregiver characteristics were associated with health status. These findings highlight the need for tailored interventions to address BPSD and support caregivers. Family-centered and system-level approaches are essential for improving outcomes for PLWD and their families.

Purpose: This study aimed to investigate the associations of disease activity, disease-related knowledge,illness perception, and coping strategies on psychological distress in patients with Crohn's disease, usinga path analysis based on the common-sense model. Methods: This was a cross-sectional study. A total of 399 adult patients were recruited from an onlinesupport community in Korea between October and December 2024. Data were collected through onlineself-reported questionnaires, using the Harvey?Bradshaw Index, the inflammatory bowel diseaseknowledge, the Korean version of the Brief Illness Perception Questionnaire, the Korean version of theCoping Strategy Indicator, and the Korean version of the Kessler Psychological Distress Scale. The datawere analyzed using SPSS/WIN 24.0 and AMOS 29.0. Path analysis was used to examine how diseaserelated knowledge and disease activity influences psychological distress through illness perceptionand coping strategies. Results: Both disease activity and disease-related knowledge were found to influence psychologicaldistress through a serial mediation effect of illness perception and avoidance coping. The avoidancecoping model showed acceptable model fit. (χ2/df = 2.52; AGFI = .96; RMSEA = .06; TLI = .91; CFI = .98;NFI = .97; IFI = .98). Disease activity and disease-related knowledge had a negative effect on avoidancecoping. The direct path from disease-related knowledge to avoidance coping showed the strongest effectwithin the research model (β=-0.29, p < .001). Significant serial indirect effect through illnessperception and avoidance coping was identified between disease activity and psychological distress(95% CI: 0.02, 0.07) and between disease-related knowledge and psychological distress (95% CI:-0.02,0.00). Conclusions: This study demonstrated the need for an integrated approach to reduce disease activityand increase disease-related knowledge as factors to effectively alleviate psychological distress in patients with Crohn's disease.

.

In this paper, we propose a soft actor?critic (SAC) algorithm with hindsight experience replay (HER), called SACHER, which is a class of deep reinforcement learning (DRL) algorithm. SAC is an off-policy model-free DRL algorithm that outperforms earlier DRL algorithms in terms of exploration and robustness. However, in SAC, maximizing the entropy-augmented objective degrades the optimality of learning outcomes. We propose SACHER to improve the learning performance of SAC. We apply SACHER to the path planning and collision avoidance control of unmanned aerial vehicles (UAVs). We demonstrate the effectiveness of SACHER in terms of the success rate, learning speed, and collision avoidance performance of UAV operation.

The performance of traditional routing protocols for mesh networks can decrease significantly in disaster recovery situations due to dynamic changes to the network environment. A number of multi-path routing protocols have been proposed to address such issues. MMQR is a multi-path routing protocol for OFDM?TDMA mesh networks which use multiple paths that can be replaced immediately by one another in case of link failures. In this paper we extend the MMQR such that the alternative paths are reordered adaptively as the communication environment changes based on the heuristic model of attractor selection mechanism by which biological entities are known to adapt to dynamically changing environment. Simulation results show that the multi-path routing protocol with attractor selection mechanism outperforms existing multi-path routing protocols in terms of packet delivery ratio, throughput and QoS.

In this paper we propose a Differentiated Services Based Admission Control and Routing Algorithm for IPv6 (ACMRA). The basic DiffServ architecture lacks an admission control mechanism, the injection of more QoS sensitive traffic into the network can cause congestion at the core of the network. Our Differentiated Services Based Admission Control and Routing Algorithm for IPv6 combines the admission control phase with the route finding phase, and our routing protocol has been designed in a way to work alongside DiffServ based networks. The Differentiated Services Based Admission Control and Routing Algorithm for IPv6 constructs label switched paths in order to provide rigorous QoS provisioning. We have conducted extensive simulations to validate the effectiveness and efficiency of our proposed admission control and routing algorithm. Simulation Results show that the Differentiated Services Based Admission Control and Routing Algorithm for IPv6 provides an excellent packet delivery ratio, reduces the control packets' overhead, and makes use of the resources present on multiple paths to the destination network, while almost each admitted flow shows compliance with its Service Level Agreement.

In this paper, a new Complete Coverage Path Planning (CCPP) scheme is proposed which combines path planning and dynamic tracking for robot operating system-based robots. For the path planning, firstly a sub-area division algorithm is considered to decompose the occupancy map according to the wall or obstacle position after simultaneous localization and mapping process. For each sub-area, an “S” shape path planning is employed, and then a Bidirectional A-star connects them. Additionally, the dynamic tracking ensures that the robot moves continuously. Simulation results show that the coverage ratio of the planning path is improved as 98% by the proposed scheme.

Knowledge Representation and Reasoning (KR & R) has become one of the promising fields of Artificial Intelligence. KR is dedicated towards representing information about the domain that can be utilized in path planning. Ontology based knowledge representation and reasoning techniques provide sophisticated knowledge about the environment for processing tasks or methods. Ontology helps in representing the knowledge about environment, events and actions that help in path planning and making robots more autonomous. Knowledge reasoning techniques can infer new conclusion and thus aids planning dynamically in a non-deterministic environment. In the initial sections, the representation of knowledge using ontology and the techniques for reasoning that could contribute in path planning are discussed in detail. In the following section, we also provide comparison of various planning domain modeling languages, ontology editors, planners and robot simulation tools.

본 연구는 해외시장 진출 스타트업의 지식학습 특성에 영향을 주는 요인을 분석하였다. 제도기반 관점과 지식기반관점의 융합적 분석을 목적으로 했다. 기존문헌 고찰을 통해 제도기반 관점 특성의 조절효과를 포함한 6개의 가설을 설정했다. 이후 설문조사를 진행하여 101개 기업을 분석 표본으로 선정했다. 구조방정식 모형인 PLS(Partial Least Square)를 사용하여 가설검정을 진행했다. 분석 결과 외부지식 특성은 모두 지식부호화역량에 긍정적 영향을 지식부호화역량 역시 스타트업 성과에 긍정적 영향을 주었다. 제도기반 특성은 외부지식 혁신성과 지식부호와 역량에 조절효과를 나타냈다. 본 연구는 제도기반 특성의 조절효과를 지식기반관점 요인과 융합적으로 통합하여 분석한데 기여도가 있다. 향후 연구에서는 제도기반 특성을 공식적, 비공식으로 나누어 분석할 필요가 있다.

This study aims to converge the analysis of institution based view and knowledge based perspectives. Based on the literature review, this study proposed six hypotheses including the moderating effect of institutional characteristics. Hypothesis testing was conducted using PLS(Partial Least Square), which is a structural equation model for 101 start-ups sample. In empirical analysis, all external knowledge characteristics have a positive effect on the knowledge codification capacity. It also positively influenced the performance. The institutional characteristics showed a moderating effects with the novelty of external knowledge and knowledge codification capacity. This study contributes to integrating and analyzing the moderating effects of institution based characteristics with the factor of knowledge based view. Future research needs to analyze the institutional characteristics in a formal and informal way.

The importance of recommendation algorithms is underscored by the advancement of AI technology and the growing demand for personalized services. However, there is a lack of empirical studies that analyze differences in algorithmic services across markets. The objective of this study is to draw attention to discrepancy in this field and develop a dual-path model to examine the factors that influence satisfaction with recommendations. By analyzing 641 respondents’ data through Partial Least Squares (PLS), it identifies differences in user attitudes towards algorithms across domains. e-commerce firms and content providers can improve personalized recommendations through the research, which highlights the importance of understanding consumer satisfaction and trust in technology to adapt to evolving AI innovations.

Purpose: This study was conducted to test the model for sleep in community-dwelling older adults. The hypothetical model based on a senescent sleep model was constructed using the multiple influencing factors of sleep and associated adverse outcomes of changes in sleep in the older adults. Methods: Data were collected from 203 community-dwelling older adults living in Korea, and analyzed using IBM SPSS 21.0 and AMOS 21.0. Results: Increased age and multiple comorbidities were associated with decreased physical activities. Decreased physical activities were associated with smaller social networks, and smaller social networks were associated with higher level of loneliness. Multiple comorbidities, decreased physical activities, larger social networks, and higher level of loneliness were associated with maladaptive sleep hygiene. Decreased physical activities, higher level loneliness, and shorter actual sleep duration were associated with worse sleep quality. Smaller social networks and shorter actual sleep duration were associated with lower cognitive functions. Smaller social networks, higher loneliness, and worse sleep quality were associated with more severe depressive symptoms. Conclusion: These findings suggest that the sleep improvement interventions not only with physical perspectives but psychosocial onesfor older adults may improve depressive symptoms as well as sleep quality.

In this paper, we present a new method to tool path planning problem for rough cut of pocket milling operations. The key idea is to formulate the tool path problem into a TSP (Travelling Salesman Problem) so that the powerful neural network approach can be effectively applied. Specifically, our method is composed of three procedures: a) discretization of the pocket area into a finite number of tool points, b) neural network approach (called SOM-Self Organizing Map) for path finding, and c) postprocessing for path smoothing and feedrate adjustment. By the neural network procedure, an efficient tool path (in the sense of path length and tool retraction) can be robustly obtained for any arbitrary shaped pockets with many islands. In the postprocessing, a) the detailed shape of the path is fine tuned by eliminating sharp corners of the path segments, and b) any cross-overs between the path segments and islands. With the determined tool path, the feedrate adjustment is finally performed for legitimate motion without requiring excessive cutting forces. The validity and powerfulness of the algorithm is demonstrated through various computer simulations and real machining.

There have been many studies regarding development of autonomous excavation system which is helpful in construction sites where repetitive jobs are necessary. Unfortunately, bucket trajectory planning was excluded from the previous studies. Since, the best use of excavator is to dig efficiently; purpose of this research was set to determine an optimized bucket trajectory in order to get best digging performance. Among infinite ways of digging any given path, criterion for either optimal or efficient bucket moves is required to be established. One method is to adopt work know-how from experienced excavator operator; However the work pattern varies from every worker to worker and it is hard to be analyzed. Thus, other than the work pattern taken from experienced operator, we developed an efficiency model to solve this problem. This paper presents a method to derive a bucket trajectory from optimization theory with empirical CLUB soil model. Path is greatly influenced by physical constraints such as geometry, excavator dimension and excavator workspace. By minimizing a energy function under these constraints, an optimal bucket trajectory could be obtained.

디지털 포렌식 분야 중 윈도우 포렌식에서 정상 파일과 의심스러운 파일을 신속하게 구분하는 것은 중요한 과제 중 하나이다. 시스템에 존재하는 방대한 양의 파일을 모두 분석하는 것은 상당한 시간과 비용이 소요되는 문제점이 있다. 본 연구에서는 자연어 처리 기법을 활용하여 파일 경로만으로 의심스러운 파일을 식별하는 방법을 제안한다. 이를 위해 윈도우의 New Technology File System (NTFS) 파일 시스템에 존재하는 모든 파일의 경로를 획득하고자 MFT(Master File Table) 파일을 사용한다. 파일 경로의 구조적 특성을 반영한 벡터로 변환하기 위해 자연어 처리에 이용되는 워드 임베딩 모델을 사용하여 정상 파일 경로를 벡터로 변환하고 벡터들을 머신러닝으로 비교하여 비정상 파일 경로를 분류한다. 실제 환경에서 수집된 데이터를 기반으로 한 실험에서 탐지 모델은 최대 94%의 정확도로 비정상 파일 경로를 효과적으로 탐지할 수 있음을 보였다.

In the field of digital forensics, particularly in Windows forensics, rapidly distinguishing between normal and suspicious files is a critical challenge. Analyzing the vast amount of files present in a system can be time-consuming and costly. This study proposes a method to identify suspicious files using natural language processing techniques based solely on file paths. To achieve this, the study utilizes the New Technology File System (NTFS) Master File Table (MFT) to obtain paths for all files in the Windows file system. The file paths are converted into vectors that capture their structural characteristics using a word embedding model, which is commonly used in natural language processing. These vectors are then compared using machine learning techniques to classify abnormal file paths. Experiments conducted with data collected from real-world environments demonstrated that the detection model can effectively identify abnormal file paths with an accuracy of up to 94%

This paper describes an advanced formation algorithm of clustered multiple robots for their navigation using flexible formation method for collision avoidance under static environment like narrow corridors. A group of clustered multiple robots finds the lowest path cost for navigation by changing its formation. The suggested flexible method of formation transforms the basic group of mobile robots into specific form when it is confronted by particular geographic feature. In addition, the proposed method suggests to choose a leader robot of the group for the obstacle avoidance and path planning. Firstly, the group of robots forms basic shapes such as triangle, square, pentagon and etc. depending on number of robots. Secondly, the closest to the target location robot is chosen as a leader robot. The chosen leader robot uses $A^*$ for reaching the goal location. The proposed approach improves autonomous formation characteristics and performance of all system.