Prompt-based fine-tuning with pause encoding reaches a maximum 95.8% accuracy for Alzheimer's detection on ADReSS transcripts, while the mean over random seeds is 87.9%.
BertForSequenceClassifica- tion
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.AS 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Leveraging Prompt Learning and Pause Encoding for Alzheimer's Disease Detection
Prompt-based fine-tuning with pause encoding reaches a maximum 95.8% accuracy for Alzheimer's detection on ADReSS transcripts, while the mean over random seeds is 87.9%.