Abstract: Embodiments afford recommendations in the accurate modeling of complex process flows. A repository is provided of known models (in graph form) of complex processes. Semantics of the repository models, are constrained within an existing vocabulary (e.g., one that does not include a particular term). During an initial training phase, a fine-tuned sequence-to-sequence language model is generated from a pre-trained language model (e.g., T5) and semantics of the known repository process models, using transfer-learning techniques (e.g., from Natural Language Processing—NLP). During runtime, an incomplete process model (also in graph form) is received having an unlabeled node. Embodiments provide a node label recommendation based upon the fine-tuned sequence-to-sequence language model. The node label that is recommended, is in a vocabulary which extends beyond the repository vocabulary (e.g., includes the particular term). In this manner, accuracy and/or flexibility of modeling of complex processes (e.g.

Abstract: A software-based process modeling tool is provided that labels and links a plurality of activities forming part of a process that is being modeled include a recommendation engine. With the modeling tool, the activities are each represented as a node in the graphical user interface. A recommendation engine is polled with information characterizing the labeled activities and their corresponding links to obtain a plurality of ranked recommendations for an unlabeled node representing a next activity in the process for selection by a user. The recommendation engine applies the information characterizing the labeled activities and their corresponding links to a plurality of rules each having a corresponding confidence value. The plurality of rules is generated using a plurality of rule templates as applied to a plurality of historical processes each comprising a plurality of labeled and linked activities.

Abstract: Computer-readable media, methods, and systems are disclosed for semantic-aware rule-based recommendation within a process modeling system in which one or more recommendations are determined based on application of one or more rules to a process model at run-time. The rules are generated from logical formulas based on certain relations within predefined process model selections, from which semantic relationships between activity labels within the process model increase the flexibility of the recommendations. Flexibility of the recommendations is further increased by applying rules such that they are fulfilled in a semantically similar sense rather than a strict sense.

Abstract: A software-based process modeling tool is provided that labels and links a plurality of activities forming part of a process that is being modeled include a recommendation engine. With the modeling tool, the activities are each represented as a node in the graphical user interface. A recommendation engine is polled with information characterizing the labeled activities and their corresponding links to obtain a plurality of ranked recommendations for an unlabeled node representing a next activity in the process for selection by a user. The recommendation engine applies the information characterizing the labeled activities and their corresponding links to a plurality of rules each having a corresponding confidence value. The plurality of rules is generated using a plurality of rule templates as applied to a plurality of historical processes each comprising a plurality of labeled and linked activities.