Dr Vivienne Ming: Mashing up tech for humans

3 Jul 2017

Dr Claire O’Connell meets with Dr Vivienne Ming (right) before the Royal Irish Academy Masterclass. Image: Royal Irish Academy

Theoretical neuroscientist, technologist and entrepreneur Dr Vivienne Ming looks to artificial intelligence and neuroprosthetics to augment humans. She spoke to Claire O’Connell.

Artificial intelligence is a hot topic. From computers that can beat chess masters or diagnose disease, to the notion of autonomous machines putting humans out of jobs, there’s a lot to think about. But, as with so many technologies, AI is just a tool, and it is how we use it that matters, according to theoretical neuroscientist and entrepreneur Dr Vivienne Ming, who was in Dublin last week.

Ming’s research and business ventures are all about ‘mashing up’ artificial intelligence, human behaviour, neuroscience and economics to augment humans and “build better people”, and this was a theme that arose repeatedly during her masterclass at the Royal Irish Academy on Thursday, 29 June.

Supported by Accenture, the masterclass brought together early-stage researchers from universities around Ireland for an informal session with Ming, during which she spoke about her career, her thoughts on AI and the basic factors that can predict outcomes in life.

‘I don’t build artificial intelligence, I build augmentive intelligence’
– DR VIVIENNE MING

Broaden discussion of AI

Ahead of the masterclass, Ming spoke to Siliconrepublic.com, pointing out that AI has cost and benefits, and she would like to see a broader discussion in the media. “When we talk about AI, we are talking about cognitive machines that are robust and adaptive and can make decisions in a fundamentally uncertain environment where there is not always a ‘right’ answer,” she said. “But far too many people are utopian or Luddite, and neither are doing us any favours.”

In particular, she believes that we get caught up in discussions about the ‘vertical innovation’ of AI, androids and autonomous robots, but there is a wealth of more ‘horizontal’ innovation where AI is enabling fields such as computer vision and the recognition of emotion. “You cannot argue against the quantum improvements [deep neural networks] have added into any field that has been structured correctly,” she said.

The more obvious robots, too, have their benefits to reduce human labour: Ming described an artificial system being developed that can distinguish crops from weeds and then take out the weeds with millimetre-level precision. “Think about it – large-scale organic farming without the need for people to be labouring in fields,” she said.

Human costs and benefits

But then there’s the human side. The IBM supercomputer Watson, which famously competed in Jeopardy!, also has the capability to analyse medical information. So will we be attending Dr Computer’s clinic in the future?

Ming isn’t convinced. “I genuinely believe it is possible to build these systems that will turn doctors into super-doctors. They can track your health. No need for lab tests. Every doctor will have [a] supply of AI tools and will be able to diagnose you and maybe even intervene before you are sick,” she said. “So why not just hire a technician to walk around with the AI and it spits out diagnoses and prognoses and it is a fraction of the cost of a doctor? But that is not how we do things economically.”

Ming’s approach puts the human element firmly at the centre of her work, which has included improving cochlear implants; building systems that can recognise emotion (and lying) in facial expressions; figuring out how to predict when someone with bipolar disorder is heading for a manic episode based on their phone data; finding the most suitable applicant for a job; hacking together a system for her son, who has type 1 diabetes, which alerts when his blood glucose is likely to dip; looking at brain electrical activity of people working in teams; and, now, through the company Ming and her wife co-founded, Socos, she is developing technology to predict educational performance in young children. “I don’t build artificial intelligence, I build augmentive intelligence,” she said.

Dr Vivienne Ming hosts a masterclass facilitated by Dr Claire O’Connell at the Royal Irish Academy

Dr Vivienne Ming hosts a masterclass facilitated by Dr Claire O’Connell. Image: Royal Irish Academy

Interesting journey

Her own career has taken an interesting road with many forks along the way, and sometimes they may not even have been road at all.

Ming grew up in California as Evan Smith, almost flunked out of high school and did flunk out of college, becoming homeless and living in her car before saving up enough money to go back to college.

When deciding what course to take, she flipped a coin to choose between economics and neuroscience. Neuroscience won and she completed the entire undergraduate degree at University of California, San Diego, in just a year.

She discovered an aptitude for programming, getting top marks in the first module she did, catching the eye of the professor running the course, and went on to do a PhD. But, despite the success, happiness eluded Ming until one night she expressed to her partner (now her wife) a wish that she wanted to be a woman, and started on a journey of gender transition. You can see more about her story in her TEDxBerkeley talk.

Benefits and taxes

While economics may have lost the coin toss for her undergraduate degree, Ming has circled back to it through her work to quantify the benefits of being able to predict outcomes for kids and intervene to improve them, and the cost of the tax on being ‘different’ in the workplace.

First, the kids: “If kids were bonds, they would be the backbone of the world economy,” said Ming. “If we did all the simple interventions that have been demonstrated in the research literature to change kids’ life outcomes, forget the human side of the story, what would it add to the economy? We looked at what if we had started 25 years ago with 50m school-age kids in the US: the real-world benefit is it would add $1.3trn to $1.8trn a year to the US economy – that is 10pc of GDP. When we ran models for South Africa, 60pc; and India, 110pc. And, it turned out, I was underestimating.”

‘If you have a female-sounding name, that’s a 40pc premium on what VCs will give’
– DR VIVIENNE MING

Meanwhile, Ming is also interested in the economic burden of being perceived as ‘different’ in the world of work, perhaps through ethnicity, gender or sexuality.

She highlighted some telling examples she found from looking at a database of jobseekers in Canada and the US, including ‘Joe v José’.

“We asked, what does José need to get the same probability of promotion as Joe [not Joseph] for the same quality of work, and, in the tech industry, we found that José needed a master’s degree or higher to be equally probable to Joe [for promotion] with no degree whatsoever,” she said. “That’s six years of tuition and six years of opportunity costs, and that adds up to $750,000.”

Nor is the news great for women: “If you have a female-sounding name, that’s a 40pc premium on what VCs will give,” said Ming, to offer an example. “So a VC will give a male $10m but the woman will get $6m.”

She also kicked back at the notion that women ‘drop out’ of the workforce, pointing out that if you have to work multiple times harder than male peers to get a promotion, then maybe the choices are fewer.

“Choice itself is fundamentally inequitably distributed,” she said. “If we doubled the income tax on men, it would be rational for men to stay home and take care of the kids. I guarantee that within 10 to 20 years, the standard story would be men just like to stay home. It is amazing how quickly these cultural norms shift.”

Predictive skills

Ming drew the masterclass to a close with insights about factors that predict life outcomes, including general cognitive ability, ‘metacognition’ (awareness and understanding of your thought processes), creativity, social and emotional skills, and purpose.

“What is predictive of whether you are a great software developer or a great salesperson? It turns out it is virtually the same thing,” she said. “Social skills, which include your ability to understand other people. [As a software engineer] you are writing code, other people need to understand it, and you have to get along with the rest of the team.”

The Royal Irish Academy Masterclass with Dr Vivienne Ming was hosted by the RIA and sponsored by Accenture. Dr Claire O’Connell facilitated the masterclass.

Want stories like this and more direct to your inbox? Sign up for Tech Trends, Silicon Republic’s weekly digest of need-to-know tech news.

Dr Claire O’Connell is a scientist-turned-writer with a PhD in cell biology and a master’s in science communication

editorial@siliconrepublic.com