Requirements Engineering for the Cognitive Enterprise

点击翻译按钮获取中文摘要

In information systems engineering, conceptual models (high-level abstractions) play an important role in enabling business architects and stakeholders to conceive of innovative solutions to business problems, serving to define requirements for system development and implementation. *** Today, a new generation of IT systems is presenting fresh challenges for requirements engineering (RE). These systems leverage the massive amounts of data from social media, sensor networks, mobile devices, and all things digital. Unlike traditional systems whose power derives from efficient and accurate execution of transactions and processes, these new software systems help generate insights and predictions. They have new capabilities and properties that are not well represented or analyzed through existing requirements models. Machine learning, artificial intelligence, and associated data science and technologies have created a modern gold rush for IT, yet organizations struggle to conceptualize their effective application in actual settings. This gap between business potential and technical possibilities suggests that new abstractions are called for at the requirements level. *** The term “cognitive computing” has been used to suggest that a new generation of IT is approaching human cognitive abilities, even surpassing humans in some tasks. From a requirements engineering standpoint, complex organizations are in fact already cognitive enterprises, regardless of the extent to which cognitive technologies are deployed. Organizations sense the environment and interpret the data, extract patterns, make predictions, formulate theories and strategies, take actions and monitor their effects, feeding back to the next iteration of the cognitive cycle. *** This proposal outlines a research agenda to develop a Requirements Engineering framework for enterprise computing based on a notion of cognitive architecture of the enterprise. Enterprise initiatives to deploy data-driven technologies such as machine learning and AI will be conceived and tested using technology-independent conceptual models, in terms of strengthening the cognitive capabilities of the enterprise to accomplish organizational objectives. The cognitive architecture of the enterprise in its current (as-is) state would be analyzed for deficiencies and baseline capabilities. The framework, which includes a visual language notation, will support and guide enterprise strategists and architects to map out and explore desirable (to-be) cognitive architectures according to the organization's maturity level. The abstract architecture will be implemented via a network of cooperating human and technological capabilities.*** Research results will include a modeling language, analysis and design methods, supporting tools, a usage methodology, as well as sample design patterns. The methodology will include guidance on mapping abstract cognitive architectures to technology implementations. **