Design and Deep Learning: Postcards From the Edge Of Artificial Intelligence

June 26, 2018

 

I’ve just facilitated an Intro To Design Thinking master class at the Artificial Intelligence in Healthcare Symposium hosted by Southern Cross University, North Coast Primary Health Network and the Centre For Healthcare Knowledge and Innovation. The master class focused on how to create the conditions for innovation through collaborative practices that emphasise a style of communication based on the principal of genuine curiosity. In the spirit of 'practice what you preach,' I arrived at the symposium early, genuinely curious to explore everything I could about Artificial Intelligence (AI) and how design might play a role in this emerging field. Here’s what I discovered.

 

Those for whom (like myself) listening to AI proselytizers requires a bit of jargon translation, I begin with differentiating AI from Machine Learning and its offspring, Deep Learning. According to Calum McClelland, blogger and Director of Projects @Leverege, “AI involves machines that can perform tasks that are characteristic of human intelligence.” This includes things like achieving repetitive tasks, identifying objects, sounds and images, and even learning.

 

Calum goes on to say that, “at its core, Machine Learning is simply a way of achieving AI.” In other words, the way a machine is able to perform human like functions is through the process of how it learns, and the way a machine learns is called Machine Learning.

 

Deep Learning is a type of Machine Learning that seeks to mimic the neurological pathways of the human brain/nervous system. One aspect of what Deep Learning engineers are trying to emulate is the transfer of learning from one domain to another. That is, the capacity to learn something on the piano lets say, and then translate that learning to a completely different life domain. For example, if I practice something on the piano and persevere until I have a breakthrough in my learning, then I can transfer that understanding to perhaps how I manage difficult relationships or other scenarios in life that require a kind of resilience/reward consequence.

According to Zac Stein, educational philosopher and co-founder of Lectica https://lecticalive.org, this kind of learning transference is a sign of a particular cognitive intelligence that sits outside of the ‘lines’ identified and valued in an IQ test. In other words, the capacity for transference of learning from one domain to another is a signal of higher orders of cognitive complexity not always recognised via conventional intelligence assessments. This kind of learning is what Machine Learning is working on… in computers.

 

Lets pause for a moment and reflect on what this might mean. On instinct, mine is a ‘glass half full’ kind of assumption; If computers can do this type of learning transference better than we can, perhaps that isn’t so bad for the evolution of our species? I also speculate that this one aspect of how humans learn is partially what makes us human, and if it can be reconstructed in non-living system, then what will ‘being human’ mean in the not too distant future? Are we the first species to design our own obsolescence? I should add that they are also working on training computers to dream. Stay with me, I’m practicing genuine curiosity.

I approached one of the presenters who spoke in depth about Deep Learning to ask some powerful follow-up questions to his talk. The most compelling for me were, “can you tell me something about the purpose behind the push to emulate types of human learning that reflect higher orders of thinking?" And, "How might this technology advance collective moral and ethical decision-making capacities in increasingly complex environments?” His response, “I don’t think Deep Learning engineers are asking these bigger questions, they just think this is the really cool next thing to discover.” I must admit, my genuinely curious self felt a bit deflated.

 

But then, a spark, a new curious, powerful question emerged: How might design play a role in developing purpose driven, ethical frameworks that help navigate this exciting new world of AI in ways that both include AND transcend the best of our humanity?

 

I’m still busy being genuinely curious with this one… For more about how to practice genuine curiosity, contact susanna@susannacarman.com

Please reload

Recent Posts

September 24, 2019

June 19, 2019

Please reload

+61 (0) 423 838 228

© Susanna Carman 2019

  • Black LinkedIn Icon
  • Twitter Basic Black