The rapid expansion of Internet of Things (IoT) devices has led to an explosion of event data, posing significant challenges for traditional process model discovery techniques in terms of scalability and discovery accuracy. These techniques rely on centralized storage and processing, which are hindered by data transfer limitations, storage capacity, and computational overhead in distributed IoT environments. Edge-based model discovery techniques offer a promising solution for analyzing large-scale IoT data. However, existing techniques suffer from low efficiency and an inability to handle complex process structures. To address these challenges, we propose EdgeIM, an efficient edge-based process model discovery technique that enhances efficiency and model accuracy. EdgeIM operates in three key stages: preprocessing and feature-preserving sampling to eliminate redundant data, local processing at edge nodes to extract key structural features, and global feature aggregation at a central node for model discovery. EdgeIM has been implemented on the open-source process mining platform PM4Py, and experimental results on nine public event logs demonstrate that, compared to existing edge-based model discovery techniques, EdgeIM significantly improves discovery efficiency while maintaining high model quality.

ABSTRACT The emotional recognition of coal mine workers across dialects faces significant challenges due to the acoustic differences between dialects. Traditional methods fail to distinguish between dialect‐specific and emotional features, leading to poor generalization. To address this, this paper proposes a recognition framework based on speech feature disentanglement, which improves the model's robustness by decoupling shared emotional features from dialect‐specific features. Specifically, the speech signal is transformed into high‐resolution time‐frequency feature maps, and a Siamese Neural Network (SNN) is used for feature disentanglement, separating emotional features into shared public feature maps across dialects and dialect/speaker‐specific private feature maps. The public encoder maximizes mutual information between same‐class samples to learn dialect‐independent common emotional representations, while the private encoder extracts dialect‐related personalized features, reducing the interference of language differences in emotion recognition. Additionally, dialect corpus information is incorporated into the task. Experimental results show that this method significantly improves emotional recognition accuracy in coal mine multilingual environments, while enhancing the system's environmental adaptability.

ABSTRACT In recent years, self‐attention mechanisms have demonstrated remarkable performance across various computer vision tasks, gradually emerging as a mainstream approach. However, compared to traditional Convolutional Neural Networks (CNNs), its high quadratic complexity and limited adaptability to 2D structures have constrained its broader application and adoption. To enhance the feature extraction capability of CNNs, this work focuses on augmenting input information and introduces a novel architectural unit called the Global Query Vector (GQ Vector). The proposed unit adopts a co‐evolutionary architecture consisting of a parallel branch and the main backbone network, which continuously integrates and refines global semantic information during forward propagation, establishing a cross‐layer, persistent context memory mechanism. This design enables progressive accumulation and refinement of contextual information, thereby enhancing the CNN's capacity to model long‐range dependencies. Building on this, we propose a novel CNN architecture named a Global Query Convolutional Network (GQConvNet). It can be seamlessly integrated into existing CNN frameworks, further enhancing their performance. For example, on the ImageNet‐1K dataset, a ResNet‐50 model augmented with GQ Vector achieves a 1.7% improvement in Top‐1 accuracy over baseline models. This work offers a fresh perspective on optimizing CNNs, with substantial academic value and practical implications.

ABSTRACT May‐Happen‐in‐Parallel (MHP) analysis serves as the basis for many concurrency bugs analyses. Inadequate handling of the coupling between locks and thread creation statements, as well as inter‐procedural locks, can lead to a loss of precision in MHP analysis. To address these issues, this paper proposes a new MHP analysis for C that operates at the segment granularity rather than individual statements. By constructing a Segmented Thread‐sensitive Control Flow Graph (STCFG) for a program, statements are grouped into different segments. Context information is added to these segments to capture the semantics of Pthreads operations, thereby identifying Happens‐Before (HB) and conflict relationships between segments. To compute MHP information for statement pairs, it is sufficient to examine the relationship between segments to infer the relationship between statements. We implement our algorithm in LLVM and evaluate it using eight test cases as well as four programs from the SPLASH2 benchmark suite. Preliminary results show that our method provides higher precision and achieves higher efficiency.

More and more business requirements are crossing organizational boundaries. There comes the cross-organization business process management, and its modeling is a complicated task. Mining a cross-organization business process aims to discover its model from a set of distributed event logs. Unfortunately, traditional process mining approaches totally neglect the privacy-preservation issue, which means the privacy of both event log and business process model. In this paper, a privacy-preservation cross-organization business process mining framework is proposed to handle its privacy issues. It includes three steps: (1) each organization discovers its private and public business process models from its event logs; (2) the trusted third-party midware takes the public process models as input and generates cooperative public process model fragments of each organization; and (3) each organization combines its private business process model with its relevant public fragments to obtain the organization-specific cross-organization cooperative business process model. To illustrate the applicability of the proposed approach, a multi-modal cross-organization transportation case is used for its validation and comparison with other methods.

Deadlock detection plays an important role in the analysis of system behavior. Several kinds of reachability trees have been proposed to analyze Petri net properties including deadlock freedom. However, existing reachability trees can only solve the deadlock detection problem of bounded or some special kinds of unbounded Petri nets. To increase the applicable scope of reachability trees in the deadlock detection field, this paper presents a new type of reachability trees called complex reachability tree (CRT). Different from others, transition sequences corresponding to root-started paths of cyclic CRTs are always firable from the initial marking. The proposed trees can completely solve the deadlock detection problem of ω-gone node-free Petri nets, but cannot guarantee to detect all the deadlocks for the other kinds of unbounded Petri nets. Their construction method and applications are presented.

Traditional process discovery techniques mine process models based upon event traces giving little consideration to workflow relevant data recorded in event logs. The neglect of such information usually leads to incorrect discovered models, especially when activities have decomposable cyclic dependencies. To address this problem, the recorded workflow relevant data and decision tree learning technique are utilized to classify cases into case clusters. Each case cluster contains causality and concurrency activity dependencies only. Then, a set of activity ordering relations are derived based on case clusters. And a synchronization-core-based process model is discovered from the ordering relations and composite cases. Finally, the discovered model is transformed to a BPMN model. The proposed approach is validated with a medical treatment process and an open event log. Meanwhile, a prototype system is presented.

Time and resource management and optimization are two important challenges for an emergency response process, by which all individuals and groups manage hazards in an effort to avoid or ameliorate the impact of disasters. Compared with a traditional business process, an emergency response process has its own features. To our best knowledge, there is no formal method to model and analyze emergency response processes by taking uncertain activity execution duration, resource quantity, and resource preparation duration into account. This paper presents such a method based on an E-Net that is a Petri net-based formal model for an emergency response process constrained by resources and uncertain durations. According to the number of available resources, execution of an E-Net is classified into the worst, delayed, and best cases. Based on a priority-activity-first strategy and corresponding algorithms, this paper finds the duration to execute each activity for the delayed case. By experiments, we prove that the proposed strategy can ensure shorter execution duration of the whole process than a conventional one. A running case of a chlorine tank explosion is given to validate the proposed method.

Recently, cross-department business processes are becoming more and more complex. Different kinds of coordination patterns exist among different departments, which make modeling and analysis work more difficult. To our best knowledge, there is no formal method to give systematic modeling and verification for the cross-department processes when considering different kinds of coordination patterns among different departments. This paper proposes such a method based on Petri nets. The WF-net model extended with resource and message factors, RM_WF_Net for short, is first introduced. Then, the formal model of tasks is proposed and its coordination relations are given. Next, RM_WF_Net modeling for intradepartment processes is investigated and cross-department coordination patterns, including message interaction pattern, resource interaction pattern, task collaboration pattern, procedure abstract, service outsourcing pattern, and process activation pattern, are formally defined. The soundness of the RM_WF_Net is verified based on the reachability graph. A running case of the cross-department medical diagnosis business process is given to validate our proposed method.

Petri nets are suitable for modeling and analysis of business processes. However, the lack of data concepts often makes Petri-net-based models excessively large and difficult to analyze especially when process logic is sensitive to changes of process attribute values. Emergency treatment processes are a typical example of this situation. To solve the aforementioned problems, this paper proposes a new hierarchical Petri net model for modeling and verification of emergency treatment processes. The hierarchical Petri net model includes three separate but closely related models, i.e., a business process logic net, a business process semantic net, and a set of case models. Business process logic nets are used to model the task dependencies disregarding semantic information. Business process semantic nets introduce data concepts to business process logic nets to model semantic information such as process attributes or conditions of sequence flows. Case models are used to model the practical routes of specific business instances. Based on the three models, a formal verification algorithm of an emergency treatment process is presented. Finally, the hierarchical modeling and verification methods are validated by an emergency treatment process of highways under snow/ice weather conditions.