Reducing Subjectivity in DI Dataset Construction

Investigate methods to reduce subjectivity in the perception, selection, and characterization of disruptive innovations (DIs) within the dataset used by the proposed predictive methodology for DI-enabled cyber cognitive attacks, for example by leveraging a team of researchers with relevant but diverse expertise.

Background

The paper builds a dataset of 11 disruptive innovations (DIs) relevant to cyber cognitive attacks and uses a combination of qualitative and quantitative models to predict future DI emergence and attributes. The DI selection and mapping rely on manual curation informed by expert survey inputs and framework mappings (MITRE ATT&CK and DISARM).

The authors acknowledge that perceptions of DI uniqueness, importance, and usage introduce subjectivity into the dataset, affecting the reliability of predictions and downstream proactive defenses. They explicitly pose an open problem to devise methods that reduce this subjectivity, suggesting diverse expert involvement as a potential approach.