Another A-Z


I’ve been a long time fan of the work that Peter Beens (@pbeens on Twitter) has done in keeping tabs of everything Google. It goes way back to a post from 2012.

Google A-Z

Sadly, Peter has had a bit of a challenge from Google itself keeping the document online. But he describes it as a labour of love and, as of the writing of this post, it’s online for all to use here.

I had to smile on behalf of Peter when I read that Google has created its own A-Z list! This list is devoted to Artificial Intelligence.

Like Peter’s list, it’s a great reference to bookmark and stay on top of.

Maybe Peter will even add it to his list of Google resources!

You can check it out at this link.

Ramona’s challenge


Yesterday, I got a challenge from Ramona Meharg that I couldn’t ignore.

I like things that are artificial intelligence-y and this seemed to have that type of approach so I gave it a shot.

The website is called Akinator.

The concept is pretty simple; sort of a digital spin to the 20 questions game that we played as kids.  When I had visited the site, there had been 700847537 games played.  That’s quite impressive.

So, ever up for the challenge, I gave it a shot.  I wanted badly to win so I chose something Canadian – hockey and just pulled a name out of my memory but then felt guilty and went with a name that might be more well known – Bobby Orr.  And, I lost.  Or rather the Akinator won.

bobbyorr

That was impressive.

How would it do with my original choice – Jean Beliveau?

How’s this for close?

JeanBeliveau

Ah, I’ve still got it!

What was interesting was the followup where I could provide details so that the next player might not be so lucky.  In other words, Akinator was learning.

Anyway, it’s 1 and 1 for me.

Are you up to the challenge?  If you’re interested in using it in the classroom, there is a child mode.  I didn’t test it extensively but none of the questions that I was asked were inappropriate.

So, thanks, Ramona.

Help with old photos


One of my favourite Adobe Photoshop Elements workshops to conduct involved demonstrating how to colourize old black and white photos.  I used to use a wedding photo of my mom and dad as my exemplar.  People were encouraged to bring in their own black and white images and work with them.

It was a great deal of fun.  But, does it still have a place in the days of artificial intelligence?  This control freak liked having control over everything!  Can I trust a program to do the task for me?

That was my challenge as I played around with this web utility.

So, first I had to grab a black and white photo – off I went to Pixabay to see if I could find something that was free to use.  This is what I chose.

architecture-3237910_640

Image by MichaelGaida on Pixabay

Next, off I went to AlgorithmIA.  A quick upload later I was ready to go.  The site even allowed me to see before and after, side by side.

colorize-compare

and the resulting image…

22051b89df022f290d1893c4f2494ede_1963d0f6-3c89-4e0e-a5ac-9d32352f8231

You’ll notice that the service stamps each of the images.

I was pleased with the results.  Of course, had I known what the original image looked like in real life in order to comment on the accuracy.

But, it’s just another example of how we might use artificial intelligence for good purposes.

Looking for someone I used to know


I still can’t find him or her.

But I’m looking.

I’m not really looking for anyone specific.  Just anyone I used to know.

Anyone who has ever walked through a shopping mall knows that there are so many different faces.  I recall once that Vicky Loras told me she saw my Doppelgänger in Switzerland.  I could swear that I saw Lisa Noble’s double in the Devonshire Mall in Windsor one day.

So, here’s my logic – such that it is.

I was inspired on this crusade using the website ThisPersonDoesNotExist.  Created by Phillip Wang, it generates lifelike human faces from an algorithm.  You can read all about it here and by following a few links followed by a few more links to get lots of details.  Plus some interesting code to read, if you’re so inclined.

Abstract: We propose an alternative generator architecture for generative adversarial networks, borrowing from style transfer literature. The new architecture leads to an automatically learned, unsupervised separation of high-level attributes (e.g., pose and identity when trained on human faces) and stochastic variation in the generated images (e.g., freckles, hair), and it enables intuitive, scale-specific control of the synthesis. The new generator improves the state-of-the-art in terms of traditional distribution quality metrics, leads to demonstrably better interpolation properties, and also better disentangles the latent factors of variation. To quantify interpolation quality and disentanglement, we propose two new, automated methods that are applicable to any generator architecture. Finally, we introduce a new, highly varied and high-quality dataset of human faces.

Selection_015

She doesn’t exist

So, all weekend, when I felt the urge, I kept whacking CTRL-R to get a new face.  I’m here to report that I haven’t found someone I used to know yet.

I can see some interesting uses for this in the classroom.   For those higher end computer science students, the reading is interesting just to see what is possible.

In terms of basic media literacy though, it presents a concrete example as to how things can be created from nothing more than a few electronic bits (and some pretty awesome programming).  It also poses an interesting inquiry to generate a face and then very closely analyse it.  Are there clues that would let you know that it’s not a real photo?

What’s Next with Robots? – Teaching Them to Lie?


My first experience with robots was with Rosie the Robot from the Jetsons. Here was the perfect servant for the home. She did everything that you would expect from a maid and never tired. She had empathy with the family and a real personality. She had her down time too, but for the most part, she was a good character in a big cast of characters and also made you think that perhaps there would come a time when a mechanical assistant would be a reality and have human traits to add value to the process. Of course, that may happen years from now.

Time passes and we see real robots in action. I recall the tour of the Daimler Chrysler plant that was given to the RCAC group. Here we see the reality as robots are put to good use. They didn’t need safety glasses, could work in difficult environmental conditions and worked with pinpoint accuracy. The ones that traversed the floor had warning horns so that you knew to get out of the road as they came along. They had their own painted pathway on the floor too. Not for them to use, but for us to know that we couldn’t.

In schools, the lead up to robotic programming is interesting and really attracts a number of students. To be able to program a turtle or a movable device is an exciting prospect for some students. Some of us long timers had long been proponents of Logo (which I still regard as one of the highlights of my own Computer Science experiences) but there’s just something so intriguing about programming something that runs around the floor and performs tasks as opposed to running around your computer screen.

No matter what, you could always count on a robot to do exactly what you told it to do. Tell it to do the wrong thing and it dutifully followed your instructions. Tell it to do the right thing and your robot would do exactly what you want. Become good enough and you and your team just might win a robotics programming competition. With some programming, they could adapt their actions but always in a predictable fashion.

Until now.

At the Laboratory of Intelligent Systems at the Swiss Federal Institute of Technology, robots now have the ability to learn about their environment and work together or they can lie to each other. http://discovermagazine.com/2008/jan/robots-evolve-and-learn-how-to-lie. What a fascinating concept. I had to read the article a couple of times to truly get a picture in my mind about what was happening.

In a robotic world where we have devices that are predictable, programmable, and with great precision — now they can be taught to lie?

Is man now becoming more precise in creating a mechanical image of himself? What next? Back stabbing, bullying, anger, …

Maybe the life of George Jetson isn’t that far off after all.

Blogged with Flock

Tags: ,