Perspectives and Views of Flexible Query Answering

The Power and Potentials of Flexible Query Answering Systems: A Critical and Comprehensive Analysis

“…The performance of DL approaches needs to be re-assessed in relation to these main situations: in the case of scarcity of training data, in highly specialized contexts, such as the legal domain [83], and for applications demanding a clear explanation of the criteria they apply in order to assess their trust and fairness [84,85].…”

“…The target user could be either an administrator of the system, who might want to control and improve the system behaviour; a scientist, who would like to know why and how a given result was obtained; or a decision maker, who needs to know if the system is compliant with regulations. For example, non-experts can find it difficult to interpret the mathematical formalization of all system processes executed to evaluate a query and to yield an answer, like indexing, matching, retrieval, and answer generation, while linguistic explanations of opaque FQAS, for example based on fuzzy decision trees, could help non-experts to understand the logic behind the retrieval or answer [85].…”

Understanding Leadership Competencies in Cyber Crisis Management: Insights from the Maersk Global Supply Chain Meltdown

“…Novel issues have to be considered, such as learning with sparse, heterogeneous, unbalanced, possibly biased and scarce labelled data. Quality assurance and assessment for Information retrieval and filtering are thus of great actuality to secure the veracity of data used for training and testing the models and the fairness of retrieval results (Viviani and Pasi, 2017;Andreasen et al, 2021).…”