Public Service Artificial Intelligence…eh?

Apr 22nd, 2015

http://bit.ly/1IzJNpS 

Funny how things come around.

On Saturday, April 18, 2015, the Toronto Star’s Kate Allen wrote a lengthy article called “The Brain Maker:  How a Toronto professor’s research took Silicon Valley by storm & shaped the future of artificial intelligence”.  It profiles Professor Geoffrey Hinton, Google intern, aged 64.

Google hired Hinton because ‘he is the godfather of a type of artificial intelligence currently shattering every ceiling in machine learning’.

(Machine learning is a computer science field dealing with algorithms that learn as humans do.)

Hinton’s specialty — neural networks — are systems modeled on the human brain.  Once a research outlier, the Star reports that in 2006 Professor Hinton made a breakthrough.  Neural networks started beating traditional Artificial Intelligence in critical tasks.

Think about recognizing speech, characterizing images, generating natural, readablesentences.

Today, neural networks, called ‘deep learning’, are one of the hottest startups in the world.

So how does AI relate to public service?

Not quite sure yet, but here’s a little story.

On March 7, 2015, I posted a post called “Why Public Servants need a SPINE” sharing a photo of my whiteboard circa May 2010.  The idea of SPINE was to create a social, semantic driven professional digital platform (sort of like LinkedIn but on major steroids without the hang ups) to foster smarter and faster collaboration, curation, crowdsourcing, content generation and cooperation between public servants in and outside of government.

In that post, I said I spoke to over 250 people, institutions & companies in the private & public sectors in Canada and the US who challenged me on the ethos, concept & operations of SPINE.  Including LinkedIn, the US Navy, DRDC,  NRCan and many others.  While testing social platforms in the process – ResearchGate, Vimeo, and others.

How does AI fit into SPINE?

One of the strongest voices challenging SPINE was Bill Howell, a chemical engineer in the Minerals & Metals Sector at Natural Resources Canada, loaned part-time to the SPINE project.  He joined me and then university student Jean-Paul Salley, along with a few others working off the side of their desks, including Jeff Kinder, to develop the idea, challenge the premise, concept and potential operations and pitfalls of SPINE platform.

A public service innovation hub before the hype.

Bill was the guy who introduced us to the world of neural networks and deep learning – one of his hobbies.

He called it “the black hole” where scientists & other assorted types globally played in the deep web with the idea that human brain networks could be algorithmically digitized to create systems that were generally intelligent. Just like us.  Machines understanding what humans say, what we mean by what is said and then take action as a result.

Spy, spooky stuff, he’d laugh with a glint in his eye.

Our small SPINE team was often pushed off kilter with Bill’s take on the SPINE concept & prototype as he injected his neural network lens into the dialogue.

He’d regularly stump me, the policy chick, who struggled to understand the science, its lexicon and ultimately its impact in lay terms in this brave new world.

The most fascinating and challenging discussion I remember having dealt with the idea of injecting the notion of ‘social’ into deep learning. Then applying it into the context of an operating social platform for public service – SPINE.

For example, can deep learning embedded in a ‘social’ technology platform like SPINE drive enhanced citizen service delivery?  What does that look like?  Can AI built on human neural networks force policy creation so as to eliminate the ‘politics’ from the evidence, debate & discourse?  Do we actually want that to happen in a democratic state?

I don’t know what the answer is, then in 2010, and now in 2015.   But I can say this, that the beauty of collaboration is just that – the messy, mindful give and take of collaborators who are willing to listen, internalize and ultimately add value to move the yardstick.

Bill and the others played on their professional margins (yet rooted in their personal passions seeing the digital futures), on SPINE.

AI in Public Service?  Its bound to be a reality in the future.

We just need to meet it head on.

And figure out the cost.

Chrystia

vision | visuals | voice mine

Note:

In 2012, www.BillHowell.ca was affected by work force adjustment.  His words:

Until recently, I was a Science Research Manager at the Mining and Mineral Sciences Laboratories of Natural Resources Canada (NRCan) in Ottawa.  My responsibility was to manage and build a new Recycling R&D program, and to find cost recovery funding for it. I have experience in corporate and government business development, market research, management and execution of scientific & engineering R&D and projects, and plant operations.

With the 2012 federal government budget cutbacks, recycle efforts in our sector (policy and science) were cut, and all employees in the science program were put on “affected” status, including myself.

At present I am looking for new responsibilities, either within or outside of the federal government.”

But what he doesn’t say, is that he moves the bar with his hobbies.

We should all be so lucky.