Can I read DiDOTS: Knowledge Distillation from Large-Language-Models for Dementia Obfuscation in Transcribed Speech on EtoBox?

DiDOTS: Knowledge Distillation from Large-Language-Models for Dementia Obfuscation in Transcribed Speech by Woszczyk, Dominika; Demetriou, Soteris is a scholarly article available to read on EtoBox.

What is DiDOTS: Knowledge Distillation from Large-Language-Models for Dementia Obfuscation in Transcribed Speech about?

Dementia is a sensitive neurocognitive disorder affecting tens of millions of people worldwide and its cases are expected to triple by 2050. Alarmingly, recent advancements in dementia classification make it possible for adversaries to violate affected individuals' privacy and infer their sensitive condition from speech transcriptions. Existing obfuscation methods in text have never been applied for dementia and depend on the availability of large labeled datasets which are challenging to collect for sensitive medical attributes. In this work, we bridge this research gap and tackle the above issues by leveraging Large-Language-Models (LLMs) with diverse prompt designs (zero-shot, few-shot, and knowledge-based) to obfuscate dementia in speech transcripts. Our evaluation shows that LLMs are more effective dementia obfuscators compared to competing methods. However, they have billions of parameters which renders them hard to train, store and share, and they are also fragile suffering from hallucination, refusal and contradiction effects among others. To further mitigate these, we propose a novel method, DiDOTS. DiDOTS distills knowledge from LLMs using a teacher-student paradigm and p

Author

Woszczyk, Dominika; Demetriou, Soteris

Published

2024

Language

EN