Artificial intelligence (AI) and machine learning (ML) are making significant inroads in solving intractable problems in healthcare, gender inequality and wildlife conservation.
The progress made by pioneers and innovators in AI/ML was highlighted at this year's Women in AI Awards Australia and New Zealand 2022 awards where Digital Nation Australia spoke to the winners and runner-ups of the most prestigious award of the night, the Innovator of the Year award, as well as the winner of the WAI Trailblazer.
AI in breast cancer screening
Dr Helen Frazer, radiologist, breast cancer clinician and clinical director at BreastScreen at St Vincent’s Hospital in Melbourne, won the WAI Innovator of the Year award, for her work in transforming women’s experience in breast cancer screening, and saving lives.
“We have curated a very large, globally unique data set for breast cancer AI research,” said Frazer.
“We've been testing and validating our models in a real-world retrospective cohort of over half a million women, and we're also working in a digital twin environment where we're prospectively testing those models in real-time, as women come into the screening pathway.”
The outcomes include higher test accuracy, due to specificity and sensitivity, as well as shortening the timeframe for women to wait for their screening results.
“An algorithm will actually pass through a mammogram almost instantaneously. Whereas for someone like me, a radiologist to read the mammogram, it takes a lot longer,” she said.
“Currently women wait up to 14 days for an all-clear result from their mammogram and 95 percent of our work in population screening of well women is normal. So there's a lot of anxiety as women wait that period to get their result.”
Frazer said she is working to implement AI to more rapidly assess images to make that turn-around the same day or within just a few days.
Radiologist talent shortages are a key challenge facing the health sector, and Frazer believes that AI can help to meet this need.
On winning the award, Frazer said she hopes that it can encourage more women and girls to work in STEM.
“We know that female participation [in STEM] is low and global figures will say 25 percent or less proportion of women are in STEM-related roles. We know also that women in artificial intelligence is even less again, and probably at best around about 15 percent,” she said.
“There are ethical, legal and social implications of machines actually making medical decisions or used to support or augment medical decisions. I believe it is so important to hear all voices, for everyone to have a seat at the table and that includes women.”
AI in genomics
Dr Denis Bauer, CSIRO’s group lead and principal research scientist in transformational bioinformatics was the first runner-up of the WAI Innovator of the Year Award for her work using AI to analyse the human genome.
According to Bauer, the human genome is made up of 3 billion letters, of which any single one can be mutated, leading to devastating diseases. Different combinations of these letters can lead to mutations of common diseases, including heart attacks and diabetes.
Bauer’s team have created a paralysed random forest machine learning implementation to analyse the huge data sets in order to identify the genes associated with increased disease risk.
“We were for the first time able to analyse the large volume of genomic data that we do have available,” she said.
“We were able to identify which locations in the genome contribute to the disease, but also how would they interact with each other, which is sort of the novel thing, which for complex diseases like cardiovascular disease, where it won't be a single gene that is actually driving the disease, this is absolutely crucial in order to understand the disease progression and then come up with new treatment.”
According to Bauer, while the technology was invented for genomics, it's agnostic in application and could be applied in other areas.
AI in the law – Preventing gender discrimination
Ramona Vijeyarasa is a human rights lawyer and a senior lecturer at the University of Technology Sydney, who was awarded the second runner-up WAI Innovator of the Year Award for her work developing the open-source tool, the Gender Legislative Index (GLI).
The GLI was developed in response to gender discrimination embedded within the law, and uses machine learning and human evaluators to determine whether a law is going to hinder or advance women’s rights.
According to Vijeyarasa, “When you take a country like Australia, we are one of the worst countries in the world when it comes to gender equality and when you think about what origination we are.
“Australia ranked 50th in the World Economic Forum’s Global Gender Gap Index last year, which is certainly not where we want it to be. To me, my work tries to contribute even in a small way to the global challenge that is gender inequality.”
She highlights discrimination in Australia’s paid parental leave scheme, which defines mothers as the primary carer, rather than that responsibility being shared by both parents. Plus laws that discriminate against single mothers as examples of discrimination existing and being exacerbated by the law.
“The machine learning is supposed to parallel human reasoning. The GLI algorithm operates as a series of ordered logical decisions based on human evaluations, which flow to a final overall score for the law. So very much following a decision tree model, but I think the AI aspect of the Gender Legislative Index is unique in that it treats all laws the same,” she said.
“It removes some of the human bias in giving an overall score for the law, which is one of the bits that makes it particularly exciting and creates a bit more integrity behind saying ‘This law meets international standards’ or ‘This law fails to meet international standards’.”
Vijeyarasa was able to put forward findings including data from the machine learning in the GLI to the Australian government when they recently called for submissions for the Workplace Gender Equality Act.
She said the same can be done for the upcoming revisions to the Modern Slavery Act.
“If the legislator is interested in advancing social justice through the laws that they're helping to enact, they can use the benchmarks in the Gender Legislative Index to say, ‘Well, have I got the ingredients to make this law a gender-responsive one?’, before that bill is put to parliament.”
AI in wildlife conservation
Camille Goldstone-Henry, founder and CEO of start-up Xylo Systems, was the winner of the WAI Trailblazer award for her work using AI in animal conservation.
Xylo Sytems uses AI and analytics to draw wildlife conservation insights from biodiversity data, which can be utilised by organisations working to save threatened species.
“Species extinctions are accelerating globally. Here in Australia we've lost more than 100 species since European colonisation and this is only getting worse,” she said.
“Now there are thousands of organisations and teams working to save our iconic species, but they don't have an easy way to connect and share information and share data to drive decision-making. This is leading to siloed and duplicated efforts, it's wasting the already finite conservation time and money that we have left to save our species.”
The organisation is using AI to aggregate data from different sources and present the data using analytics and visualisation she said.
“Once we have all of that data in the system, we know what's been done in the past and can start to predict using AI what to do in the future, particularly in the face of things like major bush fires and flooding, which we’re only going to see more and more of with climate change.”
While it currently takes seven to eight months to aggregate the right data about different species for these organisations to make informed decisions, Goldstone-Henry said that with Xylo’s technology this has been shortened to just a week.
“Aggregating data sets is done manually, usually by threatened species offices in these organisations. It's usually done by calling people and sending each other Excel spreadsheets.
"A lot of these people are spending months and months of their time getting their hands on data. We're automating all of that and we're making it faster to make wildlife conservation decisions using this data,” said Goldstone-Henry.
Xylo Systems is using drones, camera trap imaging and AI imaging to monitor species in the wild, which Goldstone-Henry believes other industries could implement in their own use cases.
“I spoke to some people who work for Toll Group, and they're interested in the way that conservation is using remote sensing because they've got trucks going across the desert and how could they implement some of those AI use cases for some of their transport and logistic use cases.”
When it comes to winning the Trailblazer award, Goldstone-Henry described the Women in AI Awards as critical in amplifying both female and Indigenous voices in the sector.
“I am of Indigenous descent, my people are the Kamilaroi people here in New South Wales and I definitely didn't consider a career in AI as a pathway for me, purely because I didn't see any women or specifically Indigenous women in this space,” she said.
“The saying is ‘you can't be what you can't see,’ so I feel Women and AI is providing that visibility for a lot of young women and young Indigenous women to consider this as a career path.”
Monash Data Futures Institute was the premier partner for the Women in AI Awards Australia and New Zealand 2022, and Digital Nation Australia was a media partner for the event.
Get the latest insights and analysis delivered to your inbox.