TIWA
Research
01 studies
-
01
Development of a lightweight interpretable model for an early-stage detection of child grooming indicators in short conversational text. This research focuses on developing a lightweight, client-side software framework that detects online child grooming in real time by analyzing conversation text directly on a minor's device before it becomes encrypted. By implementing a context-preservation mechanism that tracks interactions over time, the study aims to move away from reactive, post-hoc analysis and instead catch predatory behavior as it develops. This approach addresses critical gaps in current systems, such as outdated datasets and encryption barriers, while balancing child protection with user privacy through a text-only focus backed by parental consent.dissertation