Event Graphs Research Articles

Event Graph Neural Network for Opinion Target Classification of Microblog Comments

Opinion target classification of microblog comments is one of the most important tasks for public opinion analysis about an event. Due to the high cost of manual labeling, opinion target classification is generally considered as a weak-supervised task. This article attempts to address the opinion target classification of microblog comments through an event graph convolution network (EventGCN) in a weak-supervised manner. Specifically, we take microblog contents and comments as document nodes, and construct an event graph with three typical relationships of event microblogs, including the co-occurrence relationship of event keywords extracted from microblogs, the reply relationship of comments, and the document similarity. Finally, under the supervision of a small number of labels, both word features and comment features can be represented well to complete the classification. The experimental results on two event microblog datasets show that EventGCN can significantly improve the classification performance compared with other baseline models.

Motion Modeling and Control of Lower Limb Exoskeleton Based on Max-Plus Algebra

Max-plus algebra is a special method to describe the discrete event system. In this paper, it is introduced to describe the motion of lower limb exoskeleton. Based on the max-plus algebra and the timed event graph, the walking process of exoskeleton is modelled. The max-plus algebra approach can describe the logical sequence and safety condition in the walking process, which cannot be achieved via other conventional modelling approaches. The autonomous control of lower limb exoskeleton system is studied via the model based on max-plus algebra. In the end, an FSM (finite state machine) controller embedded with the max-plus algebra model is proposed, and the experiments show ideal speed and gait/phase period control effect, as well as the good safety and stable performance.

Causal Algebras on Chain Event Graphs with Informed Missingness for System Failure.

Graph-based causal inference has recently been successfully applied to explore system reliability and to predict failures in order to improve systems. One popular causal analysis following Pearl and Spirtes et al. to study causal relationships embedded in a system is to use a Bayesian network (BN). However, certain causal constructions that are particularly pertinent to the study of reliability are difficult to express fully through a BN. Our recent work demonstrated the flexibility of using a Chain Event Graph (CEG) instead to capture causal reasoning embedded within engineers’ reports. We demonstrated that an event tree rather than a BN could provide an alternative framework that could capture most of the causal concepts needed within this domain. In particular, a causal calculus for a specific type of intervention, called a remedial intervention, was devised on this tree-like graph. In this paper, we extend the use of this framework to show that not only remedial maintenance interventions but also interventions associated with routine maintenance can be well-defined using this alternative class of graphical model. We also show that the complexity in making inference about the potential relationships between causes and failures in a missing data situation in the domain of system reliability can be elegantly addressed using this new methodology. Causal modelling using a CEG is illustrated through examples drawn from the study of reliability of an energy distribution network.

A Temporal Event Graph Approach and Robustness Analysis for Air Transport Network

Air transport networks (ATNs) are critical to mobility and the worldwide economy. ATNs are temporal networks characterized by time-stamped information and a sequence of recurrent temporal events. Thus, static analyses disregard the real-world temporal structure of ATN. The trips are driven by considerations of cost and convenience, this way, more accurate representations, and models are necessary. Despite that, few studies have explored temporal representation for ATN. Here, we propose a temporal representation for ATN using event graphs, investigating the temporal network's topology and robustness. We also introduce a new Temporal Breadth-First-Search Pruning (T-BFS-P) algorithm to traverse the event graph paths. A study in the U.S. and Brazilian domestic ATN's robustness is presented, in which target attacks by centrality measures, giant component variation, and the robustness measure were conducted. The results show that our T-BFS-P is five times faster in execution than Dijkstra and classical BFS for the U.S. ATN. It also allows finding the shortest paths with the minimum number of nodes (airports) in the path, optimizing travelers’ connections. This work contributes to temporal network area reporting a new representation and temporal BFS algorithm via event graph, and a pioneer study to analyze robustness in event graph.

Gd-containing scintillators for thermal neutron detection via graph-based particle discrimination.

Neutron-gamma discrimination (NGD) for gadolinium-containing scintillators is a challenging issue which prevents them from being used in highly efficient neutron detectors with high signal-to-noise ratios. Pulse-shape based digital-signal processing has been selected to tackle this issue, with an emphasis on the dimensionality reduction of the raw data and capability to mitigate the influence of noise, unwanted variations, and outliers on the accuracy of classification. We present two graph-embedded non-negative matrix factorization digital classifiers in which an event graph (or a network) is introduced to find a low-dimensional structure hidden in the high-dimensional experimental data. Utilizing energy-independent normalized features of waveforms and a graph composed of mixed-source reference events, the smooth or sparse low-dimensional representation of detector signals from a Ce:Gd3Al2Ga3O12 scintillator leads to a high thermal neutron detection efficiency (77%-80%) and a high NGD ratio (neutron-gamma efficiency ratio, ∼109) simultaneously. Moreover, excellent discrimination between neutron and ambient background events has been achieved. The proposed graph-embedded algorithms may be applied not only for thermal-neutron position-sensitive detectors and single-channel detectors but also for other radiation detectors that need excellent particle (or crystal) discrimination capabilities and high detection efficiencies in many applications, including medical imaging, nondestructive testing, and security scans.

Exploiting Social Networks as a Live Mass Media Channel During Disasters for Reactions

Social networks, e.g. Twitter, have been proved to be almost real-time systems for spreading information, that provide a valuable information channel in emergencies, e.g. disasters. This paper presents a framework designed to distill actionable tweets. The framework tackles the diversity, large volume, and noise of tweets for providing users live information for quick responses. To do that, our framework first retrieves a large number of tweets to ensure the diversity. It next removes irrelevant and indirect tweets for reducing the volume, divides informative tweets into predefined classes for quick navigation, and groups tweets in a class into topics to preserve the diversity. Finally, it ranks tweets in each topic to extract important tweets for the user’s quick scan. For ranking, the framework utilizes event extraction to enrich the semantics and reduce the noise of tweets. After that, the framework builds event graphs for ranking to find out important tweets. To validate the efficiency of our framework, we took Twitter as a case study. Experimental results on five disaster datasets show that our framework achieves promising results compared to strong methods in disaster scenarios.

Failure detection and localization for timed event graphs in $(\max \limits ,+)$-algebra

In this paper, we address the problem of failure detection and localization in a Timed Discrete Event System (TDES) such \((\max \limits ,+)\)-linear system graphically modeled by a Timed Event Graph (TEG). The considered failures are changes on holding times or tokens of the TEG places that can provoke shifts between an observed outcoming timed flow and an expected outcoming timed flow (for a given incoming timed flow). Indicators are built to first detect such shifts relying on the \((\max \limits ,+)\) algebraic framework and the residuation theory. An analysis of the indicators’ values provides information about time or event failure that could have happen. Then, thanks to the knowledge of the behavior of the system through its corresponding TEG, sets of failures that could explain the detected shifts are obtained. It comes from matrices of signatures for each indicator built on each observable output of the system. An example of application is proposed to experiment exhaustively failures of type time and event on each place of the TEG.

Modeling of unsupervised knowledge graph of events based on mutual information among neighbor domains and sparse representation

Text event mining, as an indispensable method of text mining processing, has attracted the extensive attention of researchers. A modeling method for knowledge graph of events based on mutual information among neighbor domains and sparse representation is proposed in this paper, i.e. UKGE-MS. Specifically, UKGE-MS can improve the existing text mining technology's ability of understanding and discovering high-dimensional unmarked information, and solves the problems of traditional unsupervised feature selection methods, which only focus on selecting features from a global perspective and ignoring the impact of local connection of samples. Firstly, considering the influence of local information of samples in feature correlation evaluation, a feature clustering algorithm based on average neighborhood mutual information is proposed, and the feature clusters with certain event correlation are obtained; Secondly, an unsupervised feature selection method based on the high-order correlation of multi-dimensional statistical data is designed by combining the dimension reduction advantage of local linear embedding algorithm and the feature selection ability of sparse representation, so as to enhance the generalization ability of the selected feature items. Finally, the events knowledge graph is constructed by means of sparse representation and l1 norm. Extensive experiments are carried out on five real datasets and synthetic datasets, and the UKGE-MS are compared with five corresponding algorithms. The experimental results show that UKGE-MS is better than the traditional method in event clustering and feature selection, and has some advantages over other methods in text event recognition and discovery.

MVGAN: Multi-View Graph Attention Network for Social Event Detection

Social networks are critical sources for event detection thanks to the characteristics of publicity and dissemination. Unfortunately, the randomness and semantic sparsity of the social network text bring significant challenges to the event detection task. In addition to text, time is another vital element in reflecting events since events are often followed for a while. Therefore, in this article, we propose a novel method named Multi-View Graph Attention Network (MVGAN) for event detection in social networks. It enriches event semantics through both neighbor aggregation and multi-view fusion in a heterogeneous social event graph. Specifically, we first construct a heterogeneous graph by adding the hashtag to associate the isolated short texts and describe events comprehensively. Then, we learn view-specific representations of events through graph convolutional networks from the perspectives of text semantics and time distribution, respectively. Finally, we design a hashtag-based multi-view graph attention mechanism to capture the intrinsic interaction across different views and integrate the feature representations to discover events. Extensive experiments on public benchmark datasets demonstrate that MVGAN performs favorably against many state-of-the-art social network event detection algorithms. It also proves that more meaningful signals can contribute to improving the event detection effect in social networks, such as published time and hashtags.

An Asymptotic Cyclicity Analysis of Live Autonomous Timed Event Graphs

Designing a discrete event system converging to steady temporal patterns is an essential issue of a system with time window constraints. Until now, to analyze asymptotic stability, we have modeled a timed event graph’s dynamic behavior, transformed it into the matrix form of (max,+) algebra, and then constructed a precedence graph. This article’s aim is to provide a theoretical basis for analyzing the stability and cyclicity of timed event graphs without using (max,+) algebra. In this article, we propose converting one timed event graph to another with a dynamic behavior equivalent to that of the original without going through the conversion process. This paper also guarantees that the derived final timed event graph has the properties of a precedence graph. It then investigates the relationship between the properties of the derived precedence graph and that of the original timed event graph. Finally, we propose a method to analyze asymptotic cyclicity and stability for a given timed event graph by itself. The analysis this article provides makes it easy to analyze and improve asymptotic time patterns of tasks in a given discrete event system modeled with a live autonomous timed event graph such as repetitive production scheduling.

Optimization of the Time Durations by Exploiting Time Margins in Time Interval Models

Considering time interval models, which can describe a large class of models including timed event graphs and P-time event graphs, a general aim is to control the system such that it follows a one-periodic behavior starting from an initial state with a minimal or maximal cycle time. The aim of this article is the optimization of the time durations in order to achieve a given rate of production when the quantity of resources (number of pallets, machines, and so on), usually represented by the initial marking, is assumed to be a fixed datum. A prior step is the determination of the critical subsystems, whose variations can influence these optimal values and affect the obtainment of the relevant trajectories and the noncritical subsystems leading to time margins. Two approaches are proposed in that aim. A first technique is based on an adaptation of the classical Martínez and Silva’s algorithm, where each solution gives a critical subsystem, while a second approach checks each inequality of the system by an optimality verification. Using this partition of the subsystems and particularly exploiting the time margins corresponding to idle times of machines, the approach allows an optimization of the noncritical time durations when the extremum cycle times and the resources are the constant parameters of the problem. In this article, the definition of P-time event graphs is generalized by introducing the model of time supervisor place, which restricts the time behavior of a set of places. The approach is applied to a plant bakery composed of two production lines.

Imagine, Reason and Write: Visual Storytelling with Graph Knowledge and Relational Reasoning

Visual storytelling is a task of creating a short story based on photo streams. Different from visual captions, stories contain not only factual descriptions, but also imaginary concepts that do not appear in the images. In this paper, we propose a novel imagine-reason-write generation framework (IRW) for visual storytelling, inspired by the logic of humans when they write the story. First, an imagine module is leveraged to learn the imaginative storyline explicitly, improving the coherence and reasonability of the generated story. Second, we employ a reason module to fully exploit the external knowledge (commonsense knowledge base) and task-specific knowledge (scene graph and event graph) with relational reasoning method based on the storyline. In this way, we can effectively capture the most informative commonsense and visual relationships among objects in images, which enhances the diversity and informativeness of the generated story. Finally, we integrate the imaginary concepts and relational knowledge to generate human-like story based on the original semantics of images. Extensive experiments on a benchmark dataset (i.e., VIST) demonstrate that the proposed IRW framework significantly outperforms the state-of-the-art methods across multiple evaluation metrics.

Self-Supervised Prototype Representation Learning for Event-Based Corporate Profiling

Event-based corporate profiling aims to assess the evolving operational status of the corresponding corporate from its event sequence. Existing studies on corporate profiling have partially addressed the problem via (i) case-by-case empirical analysis by leveraging traditional financial methods, or (ii) the automatic profile inference by reformulating the problem into a supervised learning task. However, both approaches heavily rely on domain knowledge and are labor-intensive. More importantly, the task-specific nature of both approaches prevents the obtained corporate profiles from being applied to diversified downstream applications. To this end, in this paper, we propose a Self-Supervised Prototype Representation Learning (SePaL) framework for dynamic corporate profiling. By exploiting the topological information of an event graph and exploring self-supervised learning techniques, SePaL can obtain unified corporate representations that are robust to event noises and can be easily fine-tuned to benefit various down-stream applications with only a few annotated data. Specifically, we first infer the initial cluster distribution of noise-resistant event prototypes based on latent representations of events. Then, we construct four permutation-invariant self-supervision signals to guide the representation learning of the event prototype. In terms of applications, we exploit the learned time-evolving corporate representations for both stock price spike prediction and corporate default risk evaluation. Experimental results on two real-world corporate event datasets demonstrate the effectiveness of SePaL for these two applications.

Performance safety enforcement in stochastic event graphs against boost and slow attacks

This paper studies a performance safety enforcing problem in stochastic event graphs, a subclass of stochastic Petri net models. We assume that an intruder can attack part of the transitions to increase/decrease their firing rate such that the performance of the system violates a given safety interval. The difficulty in solving this problem is that the capability of the intruder, i.e., the number of transitions that can be simultaneously attacked, is limited. The control aim is to find a protecting policy such that the performance of the protected plant is guaranteed to be in a given safety interval. We show that this problem can be formulated as a two-player game between the intruder and the operator of the plant. By using mixed integer linear programming technique, we develop a heuristic method to compute a protecting policy that is locally optimal.

Modelling and feedback control for a class of Petri Nets with shared resources subject to strict time constraints using Max-plus algebra

We address a modelling framework and a state feedback control problem of discrete event systems with shared resources using Max-plus algebra. In this paper, we consider the class of Networked Conflicting Timed Event Graphs (NCTEGs), which are timed Petri nets with conflicts, subjected to strict temporal constraints. Firstly, an algebraic formalisation in terms of switching Max-plus linear systems is proposed to describe the dynamic behaviour of a NCTEG. Secondly, closed-loop control laws are calculated to ensure the compliance of these time constraints imposed on a certain number of places. For this, sufficient conditions for the existence of such control laws have been provided. Finally, these theoretical approaches are illustrated by a realistic example of a time-critical railway network crossing system.

Performance safety enforcement in strongly connected timed event graphs

In this paper, we tackle the performance safety enforcing problem in plants modeled by timed event graphs that are a subclass of timed Petri nets. We assume that a malicious intruder can increase the firing delay of transitions by adding delay in communication, or corrupt the transmitted data. Our target is to design a transition protecting policy to guarantee that the global performance does not drop below a given lower-bound threshold. Two classes of situations are considered: (1) the operator does not know the capability of the intruder, which means that all possible attacks must be taken into account; (2) the operator knows the upper bound of the capability of the intruder, e.g., the maximal number of transitions that can be simultaneously attacked. For the former case, we develop an algorithm using mixed integer linear programming to obtain an optimal protecting policy. For the latter situation, we develop a heuristic method to compute a protecting policy that is locally optimal.

Wafer Delay Analysis and Workload Balancing of Parallel Chambers for Dual-Armed Cluster Tools With Multiple Wafer Types

We examine a scheduling problem for a dual-armed cluster tool that processes multiple similar wafer types concurrently. It has been recently proved that the well-known swap sequence, which is widely used for single wafer type processing, also minimizes the cycle time for concurrent processing. In this article, we wish to minimize wafer delays in a process chamber, which are critical to wafer quality degradation, while maintaining the minimum cycle time. In particular, we show that concurrent processing of wafers with different processing times complicates the analysis of wafer delays significantly, and the wafer delays can be remarkably reduced by finding a proper cycle plan which is the release sequence of different wafer types. We first characterize wafer delays for a given cycle plan by analyzing the circuits of the timed event graph (TEG) model. From this, we prove that concurrent processing of wafers may cause a significant workload imbalance between parallel chambers of a process step, and hence the wafer delays increase substantially. We present that the wafer delays are minimized by a cycle plan that evenly balances workloads between parallel chambers. We also propose how wafer loading task at each process step has to be postponed to meet wafer delay constraints while maintaining the minimum cycle time. Note to Practitioners —Wafer quality control has become an essential fab operational problem in semiconductor manufacturing industry. In cluster tools, which are dominantly being used for diverse wafer fabrication stages, it has been proven that the wafer delays within process chambers have a crucial impact on the wafer quality. Accordingly, modern fabs have introduced stringent quality control to regulate wafer delays in cluster tools. In this research, we propose a scheduling strategy to minimize the wafer delays when a cluster tool concurrently processes multiple wafer types. We first show that the release sequence of different wafer types significantly impacts the wafer delays under concurrent processing, and we then propose how these wafer delays can be minimized by finding the optimal wafer release sequence.

Analysis on the Construction of Knowledge Graph of Mass Events Based on Ontology

Mass incidents pose a serious threat to public order and national security, building a knowledge base in the field of mass incidents, visualizing the relationship between entities from incidents, and scientifically reasoning about key information about incidents have become key links in effectively responding to mass violence crimes. In view of the lack of existing mass incident knowledge graph analysis methods and strong subjective reasoning methods, based on the BiLSTM-CRF model, this paper analyzes event elements and entity associations by extracting entities from the text data of mass events, the protege software is used to construct the mass event ontology model, and the visualization of the mass event knowledge graph is realized based on the neo4j graph database. The research results show that the knowledge graph established in this paper can better reflect the correlation between the elements in the mass incident, and is of great significance to the national security department to improve the emergency decision-making ability.

Analysis of P-time event graphs in (max,+) and (min,+) semirings

In this paper, we investigate the behaviour of P-time event graphs, a class of time Petri nets with non-deterministic timing of places. Our approach is based on combined linear descriptions in both (max,+) and (min,+) semirings, where lower bounds on the state vector are (max,+)-linear and upper bounds are (min,+)-linear. We present necessary and sufficient conditions for the existence of extremal (fastest and slowest) periodic trajectories that are derived from the new description. The results are illustrated by a realistic example of an electroplating process.

THE GREATEST SOLUTION IN THE INEQUALITY OF K X X LX WITH K 2 ISnn;L 2 ISnn;X 2 ISnm ARE A COMPLETE IDEMPOTENT SEMIRINGS OF INTERVAL

The greatest solution of an inequality KX X LX to solve the optimalcontrol problem for P-Temporal Event Graphs, which is to nd the optimal control thatmeets the constraints on the output and constraints imposed on the adjusted model prob-lem (the model matching problem). We give the greatest solution K X X L Xand X H with K; L;X;H matrices whose are entries in a complete idempotent semir-ings. Furthermore, the authors examine the existence of a sucient condition of theprojector in the set of solutions of inequality K X X L X with K; L;X matrixwhose entries are in the complete idempotent semiring. Projectors can be very necessaryto synthesize controllers in manufacturing systems that are constrained by constraintsand some industrial applications. The researcher then examines the requirements forthe presence of the greatest solution was called projector in the set of solutions of theinequality K X X L X with K; L;X matrices whose are entries in an completeidempotent semiring of interval. Researchers describe in more detail the proof of theproperties used to resolve the inequality K X X L X. Before that, we givethe greatest solution of the inequality KX X LX and X G with K; L;X;Gmatrices whose are entries in an complete idempotent semiring of interval