Welcome back to a new episode of Beyond the Pattern.

So today I have the great pleasure to announce Heidi Christensen.

She is a senior lecturer or associate professor you could say in computer science at the University

of Sheffield in the United Kingdom.

Her research interests are the application of AI-based voice technologies to healthcare.

In particular the detection and monitoring of people's physical and mental state introducing

verbal and non-verbal traits for the expression of emotion, anxiety, depression and neurodegenerative

conditions, like in the case of therapeutic or diagnostic settings.

So I'm very excited to present this talk here today because the presentation that she's

given has also been partially already a part of a survey talk at Interswitch.

So if you are not in the field or like me have been missing the latest progress in the

field for almost 10 years, then this is the presentation really to have a look at because

you will really get an excellent introduction into the field.

So the presentation is entitled Automatic Processing of Pathological Speech, Recent

Work and Ongoing Challenges.

Heidi, great to have you here and the stage is yours.

Excellent, thank you so much Andreas, thank you for inviting me.

I know you visited Sheffield recently and I couldn't be there so I'm glad to be able

to see you a little bit here in person.

So as you explained I'm going to talk about automatic processing of pathological speech.

I thought it might be useful to just start by explaining whereabouts I'm sitting.

So I gave a similar talk recently to John Hopkins University in the US and I thought

possibly they don't know where Sheffield is but I'm guessing there's a few more people

who know where Sheffield is.

So it's sort of more or less in the middle of England, close to Manchester.

An old industrial town but with a lovely university now and lots of modern buildings and we have

a beautiful national park just out to the west.

So if you have an opportunity do come and visit us.

So I'm going to talk about processing of pathological speech and some of the work we've been doing

in Sheffield.

I'm going to close my window because we seem to have some work happening just in a quarter.

So pathological speech in general is defined as the communication disorder where your normal

or typical speech, I've said before to call it, is disrupted.

I've mostly worked with people who have cerebral palsy, who have had a stroke or who have cognitive

impairment.

So I'll talk a little bit about each of these following on.

I think it's useful to start by just thinking about, given we're sort of engineers and computer

scientists, what is it we believe automation can contribute?

So it very often makes tasks faster and cheaper and more repeatable, more objective and I

think that's all of those are really, if you like, selling points for our work.

I think it's also important to remind everyone we work with, doctors, etc., that computers

complement the analysis that humans tend to do in routine care so we're not in a business

of replacing doctors, which of course is often a worry.

Instead we're sort of hoping to augment the healthcare provision that is happening.

So certainly in the UK there is constantly a lack of funding to the National Health Service

and so doctors are not seeing patients as much as they would like and often they have

to spend the time they see them on kind of routine tasks that were perhaps some of the

automation we could help so they could provide some of the data beforehand.

So some of the typical courses of pathological speech, as I mentioned, so cerebral palsy,

stroke and dementia or other neurodegenerative conditions.