Natural language processing for the detection of dementia
PRESENTER:
Kathleen Fraser
NRC
ABSTRACT:
Dementia is a gradual, pathological decline in cognitive ability, and
detecting dementia early in the disease trajectory is a significant
public health challenge. In this talk, I will present some of my
previous research on using natural language processing and machine
learning to automatically distinguish between short speech samples
from individuals with dementia and healthy elderly controls. I will
then discuss my recent research examining language and eye-tracking
data from people who have not yet been diagnosed with dementia, but
who have been identified as experiencing subtle changes in cognitive
function that put them at higher risk of developing dementia at some
point in the future. I will present preliminary results using this
multimodal dataset, and outline some of the issues surrounding
accuracy and interpretability of machine learning models in the
healthcare space.
BIO:
Dr. Kathleen Fraser is a researcher in the Text Analytics group at the
National Research Council. Her research focuses on using NLP and
machine learning to detect signs of cognitive or psychiatric
impairments from speech or written text, but she is broadly interested
in using NLP for healthcare applications. Fraser received her PhD in
computer science from the University of Toronto in 2016, and
subsequently completed a post-doc at the University of Gothenburg,
Sweden. She was named an MIT Rising Star in Electrical Engineering and
Computer Science in 2015, and was awarded the Governor General's Gold
Academic Medal in 2017. Fraser was also a co-founder of WinterLight
Labs, Inc., a start-up company based in Toronto.