AI can detect when your movements are off-key.
While the prospect of artificial intelligence devices monitoring one’s every movement conjures up visions of a dystopian future, your Back Page scribbler concedes there are things that AI can do which can really be quite helpful.
This is particularly the case in the field of public health, where the rapid and broadscale collection and analysis of data can really make a difference.
And the latest entrant in this burgeoning field is a doozy: a device that uses machine learning to “classify excretion events” to help detect breakouts of bowel diseases, such as cholera.
We have scientists at the Georgia Institute of Technology in the US to thank for developing a toilet-based “non-invasive microphone sensor” that can identify diarrhoea events with up to 98% accuracy without collecting any other identifiable information from the user.
Presenting their findings to the Acoustical Society of America earlier this month, the Georgia team described how they collected 350 recordings of toilet-based sounds from YouTube and a sound database called Soundsnap, covering standard defecation, diarrhoea, urination and flatulence.
The researchers then used the recordings to train an AI to recognise audible differences between the four types of excretion.
By converting the audio recordings into “spectrograms” – essentially a picture of the sound – the AI could then use an algorithm to classify each “event”. The algorithm’s performance was tested against data with and without background noises to make sure it was learning the right sound features, regardless of the sensor’s environment.
The rationale for developing this technology is to more quickly identify communal disease threats, such as cholera, allowing health officials to allocate resources and aid more effectively.
“The hope is that this sensor, which is small in footprint and non-invasive in approach, could be deployed to areas where cholera outbreaks are a persistent risk,” lead author Maia Gatlin told media.
“The sensor could also be used in disaster zones or even in nursing/hospice care facilities to automatically monitor bowel movements of patients.”
Naturally, there is also a consumer health angle down the track, where the researchers say algorithm could be linked to in-home smart devices to monitor your bowel movements and overall health.
As in: “Hey Siri, that did not sound good, what do you reckon? Can you book an appointment with my GP, stat.”
penny@medicalrepublic.com.au is covering her ears and making la-la-la sounds.
