Enterprises rely on their process-aware information system (PAIS) to conduct business. Therefore, appropriately responding to environmental changes is vital for enterprises to maintain competitiveness. However, one type of change, namely the long-tailed change (LTC), has been overlooked by traditional business process management practice because of its variety and infrequency. Just as the long-tailed effect reveals, the impact of LTC on enterprise PAIS might be no less dramatic than the impact of high-frequency changes. Since business process models are core assets of an enterprise, it is profitable to reuse them efficiently while tackling the conflict of flexibility and applicability in a timely way. This paper proposes a process model maintenance approach to responding to LTCs. By supporting business analysts to add syntax-correct annotations to existing business process models, the approach achieves an agile, error-free, and low-cost mechanism for dealing with LTCs.

To cope with the long-tailed changes, an annotation-based BPM approach has been proposed to adapts its behavior in a timely manner. It patches existing business process models rather than rebuilds models from scratch, which saves efforts and reacts to unforeseen changes quickly. However, the original annotation-based approach is at risk of improper annotations added that results in unexpected effects or leads to failure. To remedy this loophole, this paper proposes a syntax-directed annotation approach to guarantee sound reactions. we develop a scheme for designing domain specific languages based on abstract syntax trees and generating a syntax-directed editor automatically. As a result, all patched annotations on the process models are soundness guaranteed in terms of the domain specific language. Case studies demonstrate that proposed approach and tools can help domain experts to tackle long-tailed changes more easily guarantee the correct reactions.

Enterprises rely on their process-aware information system (PAIS) to conduct business. With the increasingly fierce competition, the operation environment of enterprises is constantly changing due to uncertain factors or emerging opportunities. How to react to the changes in time and make a positive response is fundamental for enterprises to maintain their core competitiveness in the new economic era. Among these changes, a kind of long-tailed change has been omitted by traditional business process management (BPM) because of its wide variety and low frequency. Just as the long-tailed effect reveals, the impact of long-tailed change on PAIS change management will be no less than the impact of high-frequency changes. On the move to digital economy, enterprises can hardly win if they cannot catch the first opportunity and deal the long-tailed changes properly. On the other hand, business process models in PAIS are core assets of an enterprise. The requirement of keeping routine processes running smoothly and reusing these assets smartly in unexpected situations runs through the whole business. To solve this problem, this paper proposes a novel maintenance approach for business process models to cope with long-tailed changes. By decorating existing process models with business-oriented annotations, the PAIS can equip with the mechanism to support original processes while react to long-tailed situations agilely with a reinforced process engine. Several experiments revels the effectiveness and generality of proposed approach for BPM in the digital economy ear.

Cluster-and-aggregate techniques such as Vector of Locally Aggregated Descriptors (VLAD), and their end-to-end discriminatively trained equivalents like NetVLAD have recently been popular for video classification and action recognition tasks. These techniques operate by assigning video frames to clusters and then representing the video by aggregating residuals of frames with respect to the mean of each cluster. Since some clusters may see very little video-specific data, these features can be noisy. In this paper, we propose a new cluster-and-aggregate method which we call smoothed Gaussian mixture model (SGMM), and its end-to-end discriminatively trained equivalent, which we call deep smoothed Gaussian mixture model (DSGMM). SGMM represents each video by the parameters of a Gaussian mixture model (GMM) trained for that video. Low-count clusters are addressed by smoothing the video-specific estimates with a universal background model (UBM) trained on a large number of videos. The primary benefit of SGMM over VLAD is smoothing which makes it less sensitive to small number of training samples. We show, through extensive experiments on the YouTube-8M classification task, that SGMM/DSGMM is consistently better than VLAD/NetVLAD by a small but statistically significant margin. We also show results using a dataset created at LinkedIn to predict if a member will watch an uploaded video.

Business process often contain many participants with different value orientations, and the fulfillment of the business value is achieved via cooperation and game among these participants in forms of services. Recently, the conceptual framework of crossover services has been constructed in pursuit of value innovation by breaking the traditional boundary and promoting convergence of service systems. Some value related factors include: 1) KPIs about activities, processes and collaborations; 2) Constraints on activities and resource utilization. The evaluation of these factors are highly dependent on participant's position, time and place of occurrence. However, in practice, such important information is not well-specified in business process models, let alone be evaluated, traced, and improved in applications. In this paper, we propose a temporal-spatial-domain ternary model to specify above information, and develop an annotation approach to attach it to standard BPMN 2.0 processes. We validate their effectiveness through case studies, and the experimental result reveals their potential advantages in pursuit of value innovation with crossover services.

In the context of a motivating study of dynamic network flow data on a large-scale e-commerce web site, we develop Bayesian models for on-line/sequential analysis for monitoring and adapting to changes reflected in node-node traffic. For large-scale networks, we customize core Bayesian time series analysis methods using dynamic generalized linear models (DGLMs). These are integrated into the context of multivariate networks using the concept of decouple/recouple that was recently introduced in multivariate time series. This method enables flexible dynamic modeling of flows on large-scale networks and exploitation of partial parallelization of analysis while maintaining coherence with an over-arching multivariate dynamic flow model. This approach is anchored in a case-study on internet data, with flows of visitors to a commercial news web site defining a long time series of node-node counts on over 56,000 node pairs. Central questions include characterizing inherent stochasticity in traffic patterns, understanding node-node interactions, adapting to dynamic changes in flows and allowing for sensitive monitoring to flag anomalies. The methodology of dynamic network DGLMs applies to many dynamic network flow studies.

Traffic flow count data in networks arise in many applications, such as automobile or aviation transportation, certain directed social network contexts, and Internet studies. Using an example of Internet browser traffic flow through site-segments of an international news website, we present Bayesian analyses of two linked classes of models which, in tandem, allow fast, scalable and interpretable Bayesian inference. We first develop flexible state-space models for streaming count data, able to adaptively characterize and quantify network dynamics efficiently in real-time. We then use these models as emulators of more structured, time-varying gravity models that allow formal dissection of network dynamics. This yields interpretable inferences on traffic flow characteristics, and on dynamics in interactions among network nodes. Bayesian monitoring theory defines a strategy for sequential model assessment and adaptation in cases when network flow data deviates from model-based predictions. Exploratory and sequential monitoring analyses of evolving traffic on a network of web site-segments in e-commerce demonstrate the utility of this coupled Bayesian emulation approach to analysis of streaming network count data.

The LinkedIn Salary product was launched in late 2016 with the goal of providing insights on compensation distribution to job seekers, so that they can make more informed decisions when discovering and assessing career opportunities. The compensation insights are provided based on data collected from LinkedIn members and aggregated in a privacy-preserving manner. Given the simultaneous desire for computing robust, reliable insights and for having insights to satisfy as many job seekers as possible, a key challenge is to reliably infer the insights at the company level when there is limited or no data at all. We propose a two-step framework that utilizes a novel, semantic representation of companies (Company2vec) and a Bayesian statistical model to address this problem. Our approach makes use of the rich information present in the LinkedIn Economic Graph, and in particular, uses the intuition that two companies are likely to be similar if employees are very likely to transition from one company to the other and vice versa. We compute embeddings for companies by analyzing the LinkedIn members' company transition data using machine learning algorithms, then compute pairwise similarities between companies based on these embeddings, and finally incorporate company similarities in the form of peer company groups as part of the proposed Bayesian statistical model to predict insights at the company level. We perform extensive validation using several different evaluation techniques, and show that we can significantly increase the coverage of insights while, in fact, even slightly improving the quality of the obtained insights. For example, we were able to compute salary insights for 35 times as many title-region-company combinations in the U.S. as compared to previous work, corresponding to 4.9 times as many monthly active users. Finally, we highlight the lessons learned from practical deployment of our system.