Under today’s natural environmental, social and technological conditions, the global and large-scale challenges faced by humanity are becoming increasingly significant, making collaborative work increasingly important. Computer-Supported Cooperative Work (CSCW) is anticipated to assume a more pivotal role in modern society and within the fabric of production and daily life. In recent years, Artificial Intelligence has witnessed remarkable advancements, and the “cooperative intelligence” of AI systems has been enhanced. The role of computers in supporting human collaboration is no longer limited to an auxiliary role but also serves as a cooperative role, bringing new challenges to the study of CSCW, ranging from basic collaborative theories to key technologies. This paper proposes a new frontier in CSCW, AI Supported Cooperative Work (AISCW) and its system research. The objective is to take use of AI technologies to enhance and optimize human collaborative activities. By delving into the interplay between CSCW and AI, this paper outlines the foundational concepts, key issues and potential research focal points of AISCW.

With the development of embodied intelligence, many studies have made progress by incorporating scene graphs and GNN into task planning. However, most methods still face challenges in fully capturing the sequential relationships between agent actions and the environment, making it difficult to handle dynamic changes and complexity inherent in embodied tasks. This paper proposes a Dynamic Graph Attention Network for Embodied Task Planning (DGETP) to process scene graph sequences and robot graphs for dynamic environment perception. In DGETP, we design a Hierarchical Dynamic Graph Attention network (H-DGAT) by employing both structural and temporal attention mechanisms to model the dynamic evolution feature of the scene. A Dual-branch Action-object Predictor (DAP) is proposed in DGETP through introducing sequences of previous actions and objects to efficiently aggregate historical information. DAP captures temporal dependencies between past and future actions through explicit sequence modeling, and reduces prediction complexity via a dual-branch architecture that separates action and object prediction while preserving their correlations through targeted feature fusion. Experiments show that DGETP improves task accuracy by over 30% in seen scenes and over 15% in unseen scenes compared to other baselines. In complex scenes, DGETP demonstrates strong generalization ability. Finally, the simulation environment indicates that DGETP achieves more goals than most of the advanced task planning method.

Deep learning on graphs, specifically graph convolutional networks (GCNs), has exhibited exceptional efficacy in the domain of recommender systems. Most GCNs have a message-passing architecture that enables nodes to aggregate information from neighbours iteratively through multiple layers. This enables GCNs to learn from higher-order information, but the model does not allow for direct captions of the local structural patterns. Our rationale is to investigate the effectiveness of capturing such local patterns for graph-based collaborative filtering to enhance model’s learning ability per layer. This technique combines lower-order and higher-order interactions during layer-wise propagation. In this paper, we propose MotifGCN to aggregate both lower-order and higher-order information in each graph convolution layer. Specifically, we develop dedicated algorithms of generating motif adjacency matrices. The matrices are then used for motif-enhanced neighbourhood aggregation in each layer. As this paper focuses on recommender systems, MotifGCN is built on the basis of bipartite graphs. Our experiments on four real-world datasets show that MotifGCN has a superior performance compared to various state-of-the-art methods.

Amid the rise of mobile technologies and Location-Based Social Networks (LBSNs), there’s an escalating demand for personalized Point-of-Interest (POI) recommendations. Especially pivotal in smart cities, these systems aim to enhance user experiences by offering location recommendations tailored to past check-ins and visited POIs. Distinguishing itself from traditional POI recommendations, the next POI approach emphasizes predicting the immediate subsequent location, factoring in both geographical attributes and temporal patterns. This approach, while promising, faces with challenges like capturing evolving user preferences and navigating data biases. The introduction of Graph Neural Networks (GNNs) brings forth a transformative solution, particularly in their ability to capture high-order dependencies between POIs, understanding deeper relationships and patterns beyond immediate connections. This survey presents a comprehensive exploration of GNN-based next POI recommendation approaches, delving into their unique characteristics, inherent challenges, and potential avenues for future research.

The rapid growth of Web services, often in the form of APIs, has made it challenging for developers to choose the right ones. While graph neural network models have improved recommendations, they struggle with complex and long-distance node relationships, limiting their effectiveness in large-scale API recommendations. This paper introduces a motif-based linearizing graph Transformer with a novel positional encoding method, integrating global and local node relationships. By leveraging motif information, we achieve linear time complexity. Experiments show our model outperforms state-of-the-art methods, highlighting the potential of global attention in Web API recommendations. The codes are available at .

In order to enable the integration between Business Process Management (BPM) and the Internet of Things (IoT), it is imperative to create an IoT service capable of encapsulating IoT device resources, converting raw data generated by IoT devices into high-level business events and transmitting these events to BPM. This would serve as a means of augmenting the intelligence and automation of business processes. Custom development of parsing and computation modules is required for different streaming data, and IoT services are built based on descriptions of business events. However, this process often involves a lot of repetitive work, and these valuable tasks are difficult to meet the requirements of changing data demands in different IoT scenarios. To solve this problem, this study proposes a code generation development framework for real-time stream data processing application, driven by business event descriptions in IoT scenarios. The framework can automatically generate real-time stream data processing applications that meet data processing needs based on domain data requirements described by business personnel in the IoT scenario. Through relevant experiments, the designed development framework demonstrates its potential to help non-software developers quickly configure Flink-based stream computing applications for scenarios and has a certain flexibility to adapt to real-time stream data processing requirements in common IoT scenarios.

With the development of IoT and AI, better knowledge and information can be learned and extracted from IoT sensory data which enables business systems to proactively provide services to customers. This paper is the first study that attempts to forecast high-level business events from raw IoT sensory event data to improve the proactivity of services and applications. We propose a deep learning based business event forecasting framework, i.e., IoT2BE, which extracts prior knowledge to identify business events from IoT sensory data, extracts features in multi-views including the spatial and temporal view, generates spatio-temporal business event embeddings, and uses a seq2seq model with attention to predict the future business events. Extensive experiments are based on two datasets including one real-world maritime ship trajectory dataset and one publicly available raw sensor dataset from a smart home environment. The results demonstrate that our framework can be effectively applied in various business scenarios.

In recent years, various Internet of Things (IoT) applications have penetrated into various areas of human life. In the IoT environment, various IoT service systems perform task collaboration by integrating with a large number of heterogeneous software and hardware devices in an open environment, providing intelligent services to the outside world. However, traditional service applications, when facing IoT with massive data, cannot meet the high real-time requirements of modern applications by relying on a pull-based response method, which also increases system operation pressure. Therefore, an event-driven architecture is adopted to provide IoT services, and service users only need to focus on how to determine the subsequent process based on event content. In this paper, we propose an IoT service interface model based on AsyncAPI, which provides a static expression of service for service subscribers, effectively expresses various operations, and makes IoT service usage more flexible. Secondly, we extend AsyncAPI to define a unified IoT service interface specification, solved the problem of protocol heterogeneity in event transmission. And complete the related code generation mechanism for the transmission of event information between systems. Finally, we apply this interface model to IoT service usage and conduct experimental verification, demonstrating the effectiveness of the IoT service interface model in event-driven communication.

In recent years, the Internet of Things (IoT) has facilitated the transformation of various traditional systems into a more convenient form of interaction and integration among people, machines, and things. In industry, there exists a significant class of process-oriented information systems, known as Process-aware Information Systems (PAIS), which are supported by business process management (BPM). The aim of business processes is to provide programming paradigms and business logic for IoT, while IoT can facilitate the quick and accurate capture of physical world changes by business processes, enabling them to respond in a timely manner. The heterogeneity in data and protocols among IoT physical entities, along with the continuous nature of their data, poses a challenge for traditional PAIS. These systems lack strict temporal semantics, and their business process event interaction mechanisms only apply to discrete events, making it difficult to efficiently handle continuous events. To address this challenge, this paper proposes a new IoT service model that has the capability to handle and respond to continuous IoT events, and enable the service to be bound to an executable element in a business process. By doing so, the IoT service can become one of the basic programming components for PAIS to respond promptly to changes in the physical world. The proposed model has been validated through case studies and experimental research, demonstrating its effectiveness and efficiency.

Proactive service is a kind of service that can be automatically provided based on logical judgment without human intervention. The traditional approach to evaluating proactive services involves conducting quantitative analyses through public datasets. However, this method can only evaluate the effectiveness of the specific algorithms and has limitations in evaluating the effectiveness of real-world applications as a whole. Determining a more appropriate evaluation approach for proactive services is a challenging problem. To address this issue, this paper proposes a method to evaluate the effectiveness of proactive services from a simulation perspective with a case of LNG (Liquefied Natural Gas) shipping as an example. Using the multi-agent simulation method, we design and implement a simulation model for LNG ships navigating at sea while avoiding dangerous areas by AnyLogic, which is capable of interacting with the deep learning model for predicting ships’ locations. Our simulation model also defines a set of metrics for evaluating the effectiveness of proactive services in the application case. We present a simulation demonstration on evading dangerous areas for LNG ships and conduct simulation experiments. The experimental results show that the simulation model can serve as an effective evaluation method for proactive services. Additionally, the simulation model can also be used to analyze differences in the application effects of proactive services implemented under different prediction algorithms.

The rapid development of web services has made it increasingly challenging for developers to find desired web services. To address this issue, researchers have developed various powerful models for service recommender systems. Recently, graph neural networks have shown promising performance in various deep learning tasks including service recommendation. This paper proposes a novel graph neural network for web service recommendation using a hierarchical attention mechanism that combines a node-level and a motif-level attention mechanisms. The node-level attention mechanism is responsible for aggregating information by the importance of different neighbors, while the motif-level attention mechanism performs a weighted combination of the node embeddings generated from different motif adjacency matrices. Finally, the generated node embeddings are optimized by the multi-layer perceptron (MLP), which in turn provide recommendations. Experimental results on real-world datasets demonstrate that our proposed model outperforms state-of-the-art approaches. Additionally, we conduct a model analysis to investigate the importance of different motifs. Overall, our proposed method shows promising performance for web service recommendation and highlights the potential of using graph neural networks in this domain.

By leveraging IoT Big Data, BPM can gain real-time physical world information to make faster and more accurate decisions, but there is a technical gap between IoT sensors and businesses. To bridge the gap, an event penetration mechanism from IoT sensors to business processes is proposed along a practical case study. This paper presents a concrete IoT-BPM application case dealing with seaborne safety ensurance in transporting liquefied natural gas (LNG), analyzes its technical challenges, and examines the feasibility and supposed effects of the BRIBOT approach.

There are a large number of real-time streaming data task processing workflows in online anomaly detection and intelligent transportation systems. Real-time data transmission is required between these processing tasks. To ensure real-time performance, it is inevitable for real-time workflow scheduling. However, due to the connected sensor data being updated in real-time, the task load of the workflow will fluctuate. In this case, the original scheduling plan is no longer optimal and needs to be renewed. Adjustments in scheduling plan bring new challenges to dynamic scheduling. At the same time, when the scheduling plan changes, the deep learning models and data used in these workflow tasks need to be redeployed at a high cost. To address these challenges, we use the data generating rate to express the task load’s fluctuation. Besides, we also consider the migration cost caused by the adjustment of the scheme in the continuous scheduling process during modeling and optimization. To achieve a low-cost scheduling scheme in a short time, we propose the NSGA-II-Seq2Seq algorithm which employs the historical scheduling plan to generate a candidate scheduling plan. The experiments are carried out in the modified WorkflowSim. Through experiment, it is found that the method proposed in this paper can adapt to the change of task load and can produce an adaptively fine-tuning scheduling scheme. In multiple consecutive scheduling experiments, the population optimization process of NSGA is accelerated, which significantly reduces the time to obtain the scheduling plan.

In this paper, we describe PACOL (Panoramic-Collaboration), a decentralized and multi-layered architecture of runtime environment for service collaboration. We illustrate by a case study why it is designed as a decentralized control architecture, and why it is designed to support multi-tenants and service solution evolution and continuous optimization. Preliminary case analysis indicates that PACOL could be a feasible proposition to realise cross-domain, reliable and optimized service collaboration for service collaboration towards Internet of Services.

Remnant subducted slab cooled the present lithosphere: Evidence from terrestrial heat flow data in eastern Fujian, SE China

In this paper, we present DataServiceHatch, a Web-based system that semi-automatically converts relational databases and stream data sources into Web Services and answers continuous queries on both traditional database tables and data streams by composing Web Services, rather than accesses databases or big data stream infrastructures directly. This can help organizations to unify the access entrance of all their data sources with just a few simple configurations and avoid exposing their data directly. DataServiceHatch also provides mechanisms to remove the need for manually writing a complex SQL-like query expression or service composition plan to answer a continuous query on data streams.

Sensor mashups provide end-users with an opportunity to aggregate and manipulate data from multiple diverse sensor data sources. A challenging problem is once the data sources are updated and propagate bottom-up to the top level, how to ensure the freshness of mashups. In this paper, an approach is proposed to build sensor mashups based on data services and to generate the synchronous policy of the mashups with good data freshness quality. The paper firstly analyzes the factors that affect the freshness of mashups, gives the problem definition based on the analysis. Then an algorithm to determine the proper synchronous policy is proposed. This paper also illustrates the proposed approach using a particular scenario.