Leadership in Improvisation

Making great teams work better together through creative collaboration.

Recently I was invited to give a talk on Leadership in Improvisation for the Peter Lougheed Leadership College at the University of Alberta breakfast event about Making Connections.

It was a delightful opportunity to share my story, and the audience was engaged, motivated and taught me some things in the process. It is my hope that by sharing the story here, the story can continue.

Note: It was critical to discuss the preparation of the talk with a sample of individuals who would be in the audience the week before, this allowed me to understand who would be in the room, what their expectations would be like, and what has worked well, and not as well, for that group in the past.

My name is Kory Mathewson; I am an improviser with Rapid Fire Theatre, and I am also a Ph.D. student in Computing Science studying artificial intelligence and machine learning at the University of Alberta.

If you want to read Part 1 of the story of me making an Artificial Improvisor, check it out here: korymathewson.com/building-an-artificial-improvisor/

Note: it was a failure. A not a total failure, because I learned a lot.

I had four short and clear goals for the morning:

  • Teach the history of improvisation and
    its growing relevance in leadership.
  • Share my stories of success and failure.
  • Distil everything I have learned over 12 years of improvisation to 6 gemstones takeaways.
  • Explore some improvisational exercises with the audience.

The same question my mom asked when she found my brother (9) and I (6) driving in a convertible we didn’t own. Let’s start with a brief history of my improvisation that should help you understand me and how we got to this point?

February 2005 at the Rapid Fire Theatre Nosebowl High School Improv Tournament. A bright-eyed memory of myself performing impeccable mime in cargo shorts, a popped collar pink polo shirt and rolled up tennis socks, with three of my best friends supporting from the bench. We lost that tournament, by a mile but we gained memories and fashion sense.

Flash forward to this month, on stage at Rapid Fire Theatre. Performing alongside the incredibly talented cast and the Mayor of Edmonton His Worship Don Iveson.

Who here knows what improvisation is? Has anyone done improvisation? Has anyone seen a comedic improvisation show? Or performed jazz? Has anyone built an ingenious solution to a problem facing scarce resources and limited time?

Improvisation, much like my life, is art and science. It sits at the intersection between the rules and pure creativity.

By the end of this talk, I promised, you will all be familiar with improvisation and the ways that the principles of improvisation can help guide your leadership and collaboration. First, it is not about being funny, it is about failing together?

I earned my stripes improvising in long 50-hour marathons. These grueling time tests stretch physical and mental abilities and provide an incomparable safe playground of practice.

Practice makes perfect; but why? Because it elucidates your patterns, obstacles and growth goals, then focuses your efforts on improvements.

The more you can adapt, response, and extemporize the more effective you will be. Veterans of the form respond proactively to sudden changes and road blocks.

The first lesson I learned in improvisation is that “the best improv performers can improvise with anyone.” They make everyone else shine; this is not necessarily true about painters, musicians, mathematicians, or lawyers, but it is right about improvisors, and more importantly: leaders.

Once I was in, I was hooked. So I did what any newly committed student of an art form does, I dug deep into the history of improvisation.

The river ran deep. Documented improvisation starts in Rome in the 4th century BC with the Atellan Farce and mask work. Flash forward thousands of years and we learn about Commedia dell-Art(e) in the early 1600’s exploring character archetypes.

Contemporaries in the field have written, studied, tested, tried and failed at shaping the art form. Each with their form of proselytizing:

  • Stanislavski: The greatest wisdom is to realize one’s lack of it.
  • Spolin 1963: Through spontaneity, we are reformed into ourselves.
  • Boal 1973: Theatre is a form of knowledge; it should and can also be a means of transforming society. Theatre can help us build our future, rather than just waiting for it.
  • Johnstone 1987: If you have a good idea, open your mouth and say something else.

And of course, no discussion of improvisation would be complete without mentioned the tortured genius of the 20th century, Del Close. A coach and mentor of many of today’s most popular comedians including Tina Fey, Amy Poehler.

With many words of wisdom, one that resonates with me as a leader is:

“Don’t bring a cathedral into a scene, bring a brick and let’s build together.” – Del Close 

Bringing us to the first gemstone takeaway.

  • Listen
  • Actually, listen.
  • Most people listen just enough to be able to respond.
  • Be willing to change.
  • Listen like this is the last thing they ever say.
  • Listen to the other is going to change your mind.

“Listen. Listen to one another like you know you are scholars. Artists. Scientists. Athletes. Musicians. Like you know you will be the ones to shape this world.” – Sarah Kay

  • Positivity.
  • Accept, agree, say “Yes”.
  • Accept and offer, say “Yes, and…” – it improves your relationships with others AND with yourself.
  • It is also a base principle of meditation. Accept your currently emotional state. Yes I am feeling like this. If this is true, what else is true.

A series of yeses takes us somewhere. All it takes is one no to stop the momentum. But what do we get when we can say: “Yes, and…”

But wait, there’s more…Practice strengthening your affirmation:

  • Practice saying “Yes, and..”
  • Use it to refocus, redirect,
    and 
    collect thoughts.
  • Improvisational leaders accept quickly and look for opportunities immediately.
  • Make your agreement prominent.

Much like building a cathedral, we can try collaborating in real time.

You can try as well, with a partner try telling a story by writing it down on a piece of paper one-line-at-a-time, for an additional challenge try writing the story one-word-at-a-time.

Then read the stories back, and see what it is like to directly collaborate on a creative piece. Remember: there are no wrong answers, no mistakes, and no judgment.

You are walking alone in a wooded forest, you have been without food and water for a few days and you are cold, scared and hungry. You come to a clearning and in the middle of the open area there is a saber-tooth tiger… you are afraid. You feel fear. Scientists think this stems from the amygdala, two almond-shaped bundles of nuclei in the temporal lobes of the brain. What is your emotional response? 

Most would say fight or flight, there is also new research on the freeze mechanism, and of course, the fourth and least favorable option in real-life, fail. You have to do something very critical in this moment, which leads us to the next gemstone.

  • Make choices.
  • In improv we often call them offers.
  • Make offers instead of asking questions.
  • Make your choices specific, unique and novel for bonus points.

Often times when we are forced with a decision we encounter a psychological phenomenon called: analysis paralysis.

Remember that you do not need to be 100% right 100% of the time. In fact, you need to be 100% right only about 10% of the time, the other times you just need to make a decision.

Your choices should be made in an attempt to make others look good. Shelve the ego and embrace the collective elevation and amplification. Endow others with power, status, and agency to create a team that functions better than individuals working independently.

I have tried on 1000 masks and understand the world slightly differently from each perspective.

Alex Williams of the New York Times has a very nice piece on friendships as we progress through life. One of the more salient points to me is that the three elements required for making close friends.

These are the exact characteristics of my interactions with improvisers all over the world. From Liverpool to Austin, to Slovenia, to New York, and back through North America… I have found my community around the world. We build relationships over space and time because we can quickly adapt, work together, collaborate openly and communicate effectively.

My research is in Reinforcement Learning. Training artificially intelligent systems to act in certain ways given rewards as feedback. This is the same way that dolphins are trained to do mind-bending stunts.

You can play this with a partner as well. Attempt to encourage your ‘dolphin’ to accomplish an unstated goal in the environment using only rewards as feedback.

This simple exercise elucidates the importance of clear communication, systems of collaboration, and shared goals.

In life,we are each living our own story (or stories), but we are major and minor characters in many other stories. We are the character that will give the right piece of advice at the wrong time or the hidden romance that ends up falling out of love and stop writing letters right before a chapter ends.

If this is the case, we should aim to make our story interesting. Aim to make offers over questions, decisions over ultimatums, and bring something specific, unique, and novel to each and every interaction.

Stories are about patterns. One of the easiest ways to understand patterns is through images. First, an action is taken establishing normal, then, with a second similar action, the pattern is established thereby creating a solid platform. Finally, the pattern is broken.

We can share the storytelling by trading back and forth on who is setting the patterns and who is breaking them. Improvisation encourages this ebb and flow, the constant back and forth of transferring energy between leading and following.

These are the characteristics of an improvisational leader who can think, speak, and act freely on their feet:

  • Listen actively
  • Amplify with positivity
  • Confidence to make authentic choices
  • Bring out the best in others
  • Tell great stories

Finally, is the failure. The most under discussed reality of the modern leader.

Here is a challenge, for the next conversation you have with a mentor or a peer that you have yet to connect on a deeper level with, ask them:

What has been your biggest failure?

Then, focus on understanding the learning that came from that failure. 

So, does anyone remember the first ‘best’ piece of advice that I ever received?

The best improvisers, make everyone else shine. I wanted to put this to the test. So I did an improv show with the audience member with the least stage experience.

Spoiler alert: He was magnificent.

So I thought, can I do it with an artificial intelligence? I would fuse my love for improvisation and machine learning. I would call it something sexy like “artificial improvisation, ” and there could be a hot Hamlet sequel skull in the bionic arm of a cyborg robot from the future.

Perfect.

And then I would do it a whole bunch, and tell everyone I knew the story and find other people around the world that were similarly passionate.

And then make an art collective in the space, and book a swath of shows in 2017. Learning and growing along the way, building businesses and research that impact millions of people.

That is how I embraced one of my biggest failures as a performer and scientist.

Questions for reflections: What do all the gemstones add up to? What is the big key takeway?

Reinforcement Learning for Profit

Is RL being used in revenue generating systems today?

Recently, one of my facebook friends, and alumni of the University of Alberta (with a PhD in Computing Science), Cosmin Paduraru posed a question:

Where is Reinforcement Learning used in revenue generating systems today?

I have been thinking about this lots over the last month as I attended two international conferences on Artificial Intelligence and Machine Learning (ICML and IJCAI) in NYC, USA. It is important to explore future prospects both inside and outside academia — In case you need a catch up, I am currently at the University of Alberta working on a PhD in Computing Science with a focus on Reinforcement Learning and Artificial Intelligence.

With the success of modern AI systems — out of the winter and into the spring — many companies have invested and continue to invested heavily into modern AI systems, backed by teams of leading researchers in the field (e.g. Facebook, Google, Microsoft, IBM, Twitter, etc.).

With that said, maybe Cosmin is right, Reinforcement Learning (Sutton and Barto 1998, and this killer-intro by the fantastically talented Andrej Karpathy) is seemingly publicly underrepresented in currently deployed systems making money in the real world, or is it?

Adapted from Sutton and Barto 1998 and WALL-E
Adapted from Sutton and Barto 1998 and WALL-E

Luckily I was at the International Joint Conference on Artificial Intelligence where I was attending a panel discussion on The Business of AI, the panel was composed of all speakers from the industry day. A desirable venue to solicit a wide variety of opinions from thought leaders in the field.

So I posed the question to them, their responses went as follows:

Peter Norvig (Director of Research at Google): “well… AlphaGo made a million bucks and then gave it away”… a recent tweet from Demis Hassabis (Google DeepMind) confirms:

Peter Stone (Founder and President, Cogitai. Professor UoT (Austin)) gave lots of great examples of recent applications:

He said,“We are on the cusp of moving from the academic lab to the industry for RL, adaptation, and lifelong learning…We are at the cusp, and that is the main motivation from Cogitai”

He also referenced work by Thomas G. Dietterich on invasive species management, wildfire suppression, by Joelle Pineau on applying RL in healthcare, and by Andrew Ng and Drew Bagnall on helicopter control. All of these could be as a practical demonstrations of specific, developing industrial applications.

Hiroaki Kitano (President & CEO SONY Computer Science Laboratories) said that this is a current research area for Sony and to expect profitability using these and advancing RL algorithms in 2-5 years. Almost 10 years after Sony’s last robotic venture, the Aibo, Sony CEO Kazuo Hirai has just recently (late June 2016) said “the robots we are developing can have emotional bonds with customers, giving them joy and becoming the objects of love”.

Guruduth Banavar (Chief Science Officer, Cognitive Computing, IBM Research) predicted that this is going to happen, sooner rather than later, and his prediction was that it will happen in the domain of conversational systems, dialog systems, and understanding the larger context of conversations. He also mentioned that the illustrious Gerald Tesauro (the man behind TD-Gammon) is working on these problems. Interesting that he did not mention Watson

Some interesting answers from industry leaders. But I was surprised that no one mentioned: recommender systems (like those on Amazon, Netflix, Yelp, and nicely formalized as an RL problem in 2005 by Shani et al.), are these systems all collaborative filtering? Surely not.

No one mentioned that Google Reinforcement Learning Architecture (here is a quick summary), which I can only imagine could be behind some of the personal recommendations and rankings that Google does behind-the-scenes on Search, YouTube, and maybe … Maps?

No one mentioned contextual bandits, sometimes called associative RL (as discussed by Li et al. 2010 for news recommendation), for serving ads and news stories. These systems are surely deployed on large-scale news sites by the publishers to maximize click-through-ratios and create a personalized experience. Microsoft recently announced Multiworld Testing Decision Service, for making context based decisions… I guess there were no Microsoft representatives on the panel to toot this horn (thanks for the catch Pardis)

With so much potentially out there, why was there no mention of these use cases for reinforcement learning? Where else could RL be hiding in the money-making wild? RL seems like an ideal candidate for systems of personalization on large-scale, sequential decision-making problems… so what am I missing?

Building an Artificial Improvisor

In June 2015, I started a journey to develop an Artificial Intelligence that I could perform improvisational theatre with. April 8, 2016, my dream becomes a reality. This post documents the progress…

Bonfire_AI

In June 2015, I started a journey to develop an Artificial Intelligence that I could perform improvisational theatre with. April 8, 2016, my dream becomes a reality.

Background

As you may or may not know, I am an improvisor, and a Ph.D. student in Computer Science, studying artificial intelligence. Early on in my improv career a very, very good improvisor told me:

A good improvisor looks great on stage. The best improvisors make everyone else look good.

This was advice that I have held closely through my artistic journeys. Lee White (twitter, wikipedia), a talented Canadian performer (and my improv Uncle) told me about a show where he selected from the audience the individual with the least stage / public speaking experience and then did a show with them. I was up for the challenge and pitched the show to then Rapid Fire Theatre’s Artistic Director Amy Shostak. She fostered the idea (as all great improvisors do) and I performed the show multiple times to great success.

1511359_10151947650546582_7070474645998644371_n

If I could do improvisation with anyone in the audience, could I improvise with anything? (this reminds my of another mentor and early guide Jacob Banigan (twitter), and his work in solo improvisation).

Some more cool background research that I found and followed up on:

The Journey to AI (artificial improvisation)

Improvaganza 2015

It was at Improvaganza 2015 that the idea of building an artificial intelligence came to be. Sitting at the pub with friend-from-across-the-pond Adam Meggido, who just recently won an Olivier (like a TONY for Brits) for Showstopper on the West End of London.

Adam told me to embrace my scientific side, he challenged me to bring the best in artificial improvisational technology to the stage. Most importantly, he told me in no uncertain terms:

The first time will be the worst time, but the science is getting better every year — every month even. So, just start!

Adam was a great initial sounding board, he helped me wrap my ideas around the basic foundations of why this kind of project would happen.

How and why science and art intersect?

Bringing AI to the stage is fascinating, but for me it was due to three main points:

  1. Science broadens our artistic understanding;
  2. Art broadens our scientific curiosity; and
  3. Both can be used to explore and understand humanity.

Deep, I know, but it is in these points that I found my main inspiration.

Science broadens Art

The worlds of art and science are intrinsically linked. We, as artists, grow experimentally with every performance. Similarly, we as scientists, are imbued with creativity in hypothesizing. The combination of technology and storytelling is not the novel convention. Differently stories are told through different means. The medium is the message (as Marshall McLuhan says).  The way that we tell the story, symbolically connects to the messages being portrayed in the contained narrative.

We are connecting the minds of the audience with the performers, sharing their awareness of the performance live and in real-time with the performance as it is happening. By allowing the transparency between the story-teller and the audience, we can collaborate the share in the experience. Our art will push our technology, but our technology should push our art as well. As audiences evolve they want to be challenged, and we should not be afraid to experiment with the newest and most powerful tools to capture that excitement.

Art Broadens Science

Creatives are paving a road of technological experimentation. By fusing the worlds of natural and artificial we can start to see the links between the sentient and the synthetic. We are using a means that is appreciated, understandable and approachable in performance theatre, to share knowledge of advancing technology.

Artificial Intelligence is a tool, this is a specific implementation of that tool. It is the medium with which we can tell these stories.  Emotional Artificial Intelligence (that is, one that can portray empathy or sympathy or compassion) is somewhat abstract and unappreciated in the scientific literature. By creating these links we are able to somehow add dimension to the discussion.

Using Both to Understand Humanity

What is it to be human? What are basic human values? Is morality relative? These are all extremely difficult questions to think about, let alone to start to answer. Perhaps, the combination of the human and the machine, raw and live in front of an audience which shaping the experience, can help to understand how to approach answers.

Can a robot perform a Shakespearian monologue? Can a robot act? Can a synthetic voice perform all the parts in a play? Can the robot be the actor and the director and the audience? Where does the human NEED to come into a theatre piece, if at all? Theatre helps us to understand the human condition, but perhaps it could help us to understand the human as well.

Performance Ideas Pt. 1

The ‘Adam’ discussions (the creationist metaphor is not lost on me) also lead to two interesting ideas for the performance:

  • A potential branching cue structure where I can cue the booth to play the next sound cues, which could be the voice of the AI. Rather than the deterministically selected cue, it could choose randomly from a set. This selection could also be biased from choices it has (or I have) made in the past.
  • The idea of the AI bring a voice that convinces of how real it is by physically embodying an audience member. Perhaps the voice helps select someone from the audience and they then are the physical embodiment of the voice.

January 2016

It was deep in the cold Edmonton winter when the Reinforcement Learning and Artificial Intelligence lab at the University of Alberta was visited by Dr. Michael Littman. A renowned AI scientist, who made time to sit down and chat with me about research ideas and … creative projects. While at I was at first hesitant to share some of my more out-there ideas, Dr. Littman embraced the ideas wholeheartedly. He imbues the sense of “Yes, and”, a notable quality for an academic.

Dr. Littman was thrilled with the idea of the show, and connected me to a fiend of his named Tom Sgouros. While I didn’t realize it at the time, this was one of the critical connections between the art and the science.

Tom Sgouros and JUDY

judy

In Spring 2005 (yes, that is the correct date, I know it seems like a long time ago for what I am about to tell you, so see the date on the headline of this article), Tom performed alongside a robot that he built named Judy. Tom explored the question:

IF YOU BUILD A ROBOT SMART ENOUGH TO DO THE DISHES, WILL IT ALSO BE SMART ENOUGH TO FIND THE TASK BORING? FIND OUT IN THIS EXCLUSIVE INTERVIEW WITH A ROBOT.

Needless to say, this was utterly inspiring. Tom told me about his history of performance, with so many inspiring shows (and creative uses of VCRs, fake tapes, and cue buttons intelligently hidden around the stage) .

Performance Ideas Pt. 2

I told Tom about the show I was imagining:

  • The show is somewhere in between Her, Pygmalion and tomorrow’s tomorrow. It exists in a world where humans and robots can have relationships. It is, in that sense, some what futurism. It provides some commentary on the judgement of others on the choices we make. I had the idea of using the robot as an introduction to the surrogate; can the robot convince an audience member to be the vessel, and at that point, it could be that the human is the intelligent being?
  • This could show the tradeoff between the human and the machine. Plus then I would get to have a lovely, improvised scene, with someone from the audience, as they are ‘acting’, as if the AI gave them full control.
  • Perhaps there is a place for audience interaction, or maybe the audience has some driving power over the intelligence, how can this feedback help to drive the show.
  • Perhaps there is a natural language processing piece to incorporate, or a dynamic voice synthesis piece (as those are still very developmental).

The Hard Part

With my background research done.  It was time to start with the implementation

My name is  Pyggy

I wanted a robot that I could improvise with, but I was less focused on the hardware and more interested in the dialog engine. With improvisation listening is key, so I needed an intelligence that could hear me and speak. I started building Pyggy.

The name Pyggy comes from the name Pygmalion, and the fact that Pyggy is build in Python (see what I did there), it also stems from this haunting video:

Seemingly alone, Pygmalion sought to create for himself a perfect, pure, unsullied companion. He used his particular skills to this end: he created a statue bride.

What I was setting out to do was build a perfect improvising companion, and there are eternal dangers of seeking idealist ideas.

Pygmalion is also the name of a George Bernad Shaw play about an academic who makes a bet that he can train Cockney flower girl Eliza Doolittle to pass for a duchess. How fitting. Ps. if you haven’t yet, turn on My Fair Lady while you read the rest of this post. You won’t miss much, just the credits and some exposition.

The Technology

It was less the fancy robotics that I wanted to focus on, in improv you have to play so many characters, so fleshing out a body isn’t the top priority. So, having made the decision to stick with the software implementation I started  programming.

Pyggy is free and open-source if you want to try to get it running on your own system.

Behind Pyggy there are several modules:

  • Pandorabots Chatbot
  • Chatterbot learning Chatbot
  • Speech recognition
  • Speech synthesis

Speech Recognition and Synthesis

Speech recognition is done by the python module of the same name. Pyggy uses the newly released Google Speech API (which was majorly updated during development, see Adam, technology won’t stop progressing, so it was right to jump in before it was ‘ready’).

Synthesis is done by the onboard speech text-to-speech system for the mac, but I did download some better sounding voices (for my ears, the best ones are Tom and Samantha).

Pandorabots Chatbot

The Pandorabots chatbot playground is a great place to get your feet wet starting to play with chatbots. They have some great AIML libraries which serve as a nice structured foundation for their bots. As well, their API from chatbots.io makes it super easy to start your chatbot up and deploy it live.

But their training is quite difficult as it all relies on the mostly deterministic AIML. It made for some interesting conversations, but I couldn’t really train it deeply on conversational knowledge, and it worked best only when I was asking the ‘right’ questions.

Finally, it is not free and not open source.

Chatterbot Chatbot

The Chatterbot Chatbot from  Gunther Cox is free and open-source. Which is great because while it is a strong code base to start a project on, I found that I wanted to train it on huge datasets and it was quite slow.

I was able to fork the repository and modify things nice and quickly so my training could happen lightning quick. Maybe it will even make it into the main codebase!

I was also able to deploy it quickly on a working training interface (which is currently not running, but as I wanted to freeze the training), so that I could enlist the help of my fellow improvisors to help seed and train the model.

Screen Shot 2016-04-08 at 3.23.10 PM

The chatbot runs on a database (MongoDB) which is trained on a huge corpus (Cornell Movie-Dialogs Corpus) of conversations.

This corpus is gigantic (300k utterances from over 10k characters), due to the training speed limitations of chatterbot I had to settle for training on a random subset of 20000 conversations, and it still took an hour to train. For the next iteration, I want to use the full corpus.

It works by fuzzy matching my speech recognized input string with some strings in its dictionary and then producing the most likely response to that input.

Gluing it Together

I combined the chatbots, to get the best of both worlds. The chatbot-like feel of the Pandora bot, and the deep conversational knowledge and interest of the chatterbot, and I got Pyggy.

Visualization

As there will be an audience for the performance, and it will be taking place in a well-equipped theatre, I though that it would be best to find an interesting visualization for Pyggy.

A great recommendation at the 11th hour led me to build a nice face with talking animation using the Magic Music Visualizer and Soundflower to port the audio through.

Performance Ideas Pt. 3 – You’re On

I have been so incredibly supported on my academic and creative artistic sides, it has been a great learning journey over the last year or so.

So, April 8 is the performance. I am performing alongside Pyggy in Rapid Fire Theatre’s Bonfire Festival. Thanks to the current Artistic Director Matt Schuurman for giving me the chance to do it, and pushing me when things were faltering.

I plan to wear a headset microphone, and project Pyggy on the screen, and have a scene with Pyggy. It sounds weird to say, but it is going to happen.

I will pray to the non-denominational demo gods that things don’t crash, I will break legs with the other actors on stage. It feels like my worlds of science and improvisation are colliding right in front of my eyes.

Or, your eyes, if you are going to be there… This time.

But like Adam says, the first time will be the worst time, because technology is only going to progress, Pyggy is only going to get better at conversing, and I will only get better at training Pyggy.

unnamed

Thank You

As I reflect on Pyggy, I am overwhelmed with the passion and enthusiasm of friends, colleagues, and collaborators. It was common for me to bring it up and for friends to quickly offer help and ideas on creation, so thank you to: Paul, Sarah, Matt, Joel, Brendan, Paul, Joe, Nikki, Leif, Julian, Paul, Lana, Stu… it is moments like this that helped me through the ‘this is never going to work’ phase:

Screen Shot 2016-04-08 at 3.38.26 PM

 

 

 

Visiting Washington State University

Travelling to Washington State University to guest lecture and speak on the Autonomous Anatomy we are building at the University of Alberta. Sometimes you have to travel to new places to see what is really happening here at home.

Recently, I was invited to perform TEDxRFT at the Seattle Festival of Improvised Theatre. It was a delightful opportunity to do our show for a new audience in a new city, to travel to Washington State and to visit some old friends from Portland and abroad!

20160229-visit-slides.001

It was also an opportunity for me to take a shot in the dark and send an email to a professor at Washington State University whose work I admire very much. Dr. Matt Taylor works in the domain of reinforcement learning in the Intelligent Robot Learning Laboratory, or IRL at WSU. I am familiar with his work through my research on current topics in the field of teaching robots (see a recent paper of our thought framework).

20160229-visit-slides.002

The IRL is an amazing place with research in the areas of Reinforcement Learning, Transfer Learning in heterogenous domains, Multi-agent Exploration and Optimization, Autonomous robots, and Deep Knowledge Transfer… read more about their current projects.

Matt enthusiastically invited me to come visit the lab at WSU, to give a talk as part of the Smart Environments Research Center Distinguished Speaker Series and to guest lecture in his Introduction to Robotics class (I gave a lecture titled Don’t Shoot the Robot, inspired in large part by the dog training work of Karen Pryor).

There are many things happening in and around the lab that I very much enjoy:

  • Interest in project teams (3D printed prosthetic hands controlled by EEG and RoboSub);
  • Hardware Hackathon to bring together small, diverse teams for a short hack on a new project;
  • Great science and engineering on remote biophysical monitoring, and assisted catheter insertions;
  • Lots of access to robotic development hardware (UAVs, drones, Turtlebots); and
  • Great science on interactive reinforcement learning (like this amazing work).

Lots of this work happens in the Frank Innovation Zone, or Fiz, which is an amazing Community Studio with Wood, Desktop and Metal Fabrication, Electrical Testing and Fabrication and lab space to experiment with the new robots and algorithms. Students are trained on the tools and machines and they are further supported by staff who can purchase raw materials and advise on projects.

20160229-visit-slides.014

It is fun to  connect with amazing scientists and engineers who are physically far away, but, when it comes to the research we are doing, it feels like they are right next door.

I can’t stress enough how important it is to get out of the place you are working in, and to travel to gain exposure to the work happening elsewhere in parallel. It also helps give you perspective on the work you are doing and how it fits into the global knowledge landscape.

Special thanks to the amazing students of Matt’s who made my stay so easy and enjoyable: Gabe, Bei, James, Chris and everyone else who shared with me their research ideas, goals and progress.

IMG_1664 2

Applying for and Winning Major Canadian Scholarships

TL;DR I present my secrets to applying for and winning major scholarships.

Recently (September 14 2015) I was invited to present what I knew about applying for and winning major scholarships to the incoming graduate students in Computing Science at the University of Alberta.

This is the talk that I wish I heard ten years ago when I was just starting out as an undergraduate student.

TL;DR: Download my slides directly and get right into them. If you are going to use anything from the slides please be sure to properly attribute it.

Preface

I think it is critical for new students (Graduate and Undergraduate) to hear directly from students who have won major scholarships/grants/awards to get a sense of what those students did right, and how they can best make their application competitive.

There are many points that students will tell you (which you will read below), that officials at the University, or the granting agency, can not talk about.

While these keys are specifically written for major scholarships (requiring in-depth prepared sections academic and extra-curricular achievement) much of the information is more generally relevant.

Major Scholarships in Science and Engineering

About Me

To date I have secured almost $100,000 in scholarships, grants, and awards.
My academic history looks like this:

  • 2010 – B.Sc. Electrical (Biomedical) Engineering (Photoacoustic microscopy)
  • 2014 – M.Sc. Biomedical Engineering (Measuring blood flow and oxygenation with MRI)
  • Current – Ph.D. in Computing Science (Artificial intelligence in prosthetic limbs)
from Prosthetic Devices as Goal-Seeking Agents. Pilarski, Sutton and Mathewson. 2015.
from Prosthetic Devices as Goal-Seeking Agents. Pilarski, Sutton and Mathewson. 2015.

Inspirational Quote

Someone has to win, convince them it is you. — Kyle Mathewson

10 Keys to Applying and Winning

  • Find all Award Details
  • Follow a Schedule
  • Follow the Rules
  • Use Available Resources
  • Don’t Waste Words
  • Trust your Past Work
  • Secure Excellent References
  • To Win More, Apply More
  • Review, Review, Review
  • Do Something Today

Find Award Details

Locate all the information on the relevant scholarships. Collect it in a quick easy to read manner. Your department, or faculty, will most like have a repository like the University of Alberta does here. Go thank them in person for doing the hard work of collecting all the information. Your gratitude will go a long way.

Follow a Schedule

vertical_holdStart right now. Open your calendar.
Set internal deadlines, and stick to them.
Signatures and references take time.
Your countdown should start the moment  the information is posted.
Add award reviews to your daily, weekly, monthly, and semester check-ins.

My Usual Schedule

  • 12 months before: Research past posting
  • 6 months before: Prepare required materials
  • 2 months before: Compile all required materials
  • 1 month before: Final version for peer-review
  • 2 weeks before: Proofread final version
  • 1 week before: Make final hard copies
  • 2 days before: Submit

Follow The Rules

color_inside_the_lines

Agencies try to make it easy.
The best place to look for how to fill the form out is to find the application information.
Note word limits, page limits, and formatting.
One Canadian granting agency reported 25% of applications are incomplete (The Peer Review).

Use Available Resources

naturalresources580_1

Connect with friends and colleagues about awards.
Contact past winners to discuss what worked for them.
Engage with potential supervisors and discuss framing a research proposal.
Make friends with just enough English majors, and scratch their backs.
If you don’t have English major friends, contact a writing support group like the Centre for Writers.
If you want more administrative support, contact the Faculty of Graduate Studies and Research (FGSR).
Contact your Graduate Advisor, they are an incredible resource for departmental needs.

Research Past Winners

Find past information on agency websites.
Investigate the work past winners did by cross referencing with Google Scholar and Google.
How did they present their work in a compelling way? What was their hook?
Read past applications.

Ask Questions to the Agency

If something on the application is unclear, it is better to clear it up early (prior to application) with someone at the agency.
Keep moving until you get the answer, be flexible.
Added benefit is that it then creates a connection with the agency, someone who may help you in the future.
Then they get a chance to help you before you even apply.

Don’t Waste Words

Every application you create, creates value.
Don’t waste words. Past applications and award descriptions are gold mines for personal statements, and proposals.
Write timelessly. Reference time directly rather than “recently” or “last year”.
Keep your entire history on a professional internet presence. Forces you to stay updated.
Review background literature (with reference) on interesting research questions.

Clearly Describe Your Research Proposal

You should be able to clearly describe your work in 1500 words, or 500, or 150, or 15, or 5 words.
Describe your work out loud. Record it.
Note the questions people ask about your work. These are the points that you can be more clear about next time.
Your proposal does not have to be your current work, or all of what you plan to work on.
It does need to be clear, concise and contained.
Be honest with what you can accomplish. 

Know Your Audience

Use appropriate language.
Often science-minded individuals working outside of your specific field review your application.
Tailor you application to the funding agency.
You should be able to explain it to me, or the person next to you, or my grandmother.
“Good writing will not save bad ideas, but bad writing will kill good ones” – Kracier, The Art Of Grantsmanship

Trust Your Past Work

71WupPBFm2L._SL1036_

My past research is full of challenging research (Scoliosis brace internal biomechanic monitoring, EEG seizure detection).
Failure teaches more than success.
Everything you do can be described as significant, even if you think it’s not. Talk to someone about how to expand your experiences.
Your grades are sufficient.
Your professional and volunteer experience is suitable and character developing.
Your research is worthy of funding.

Include all Relevant Experience

Flickr_IntroverLeadersCommunicationMED-kenfagerdotcom4886342418_db75a10808

Your past Academic, Research, Corporate, Startup experience is relevant.
Highlight communication and leadership. These two skill-sets are highly regarded.

Develop Yourself

Actively attempt to fill the gaps in your character development.
Seek out volunteer and/or internship opportunities.
Maintain or improve your grades.
Get involved in campus culture. If what you want doesn’t exist yet, make it happen.

Secure Excellent Reference Letters

Find someone who has:

  • Worked with you closely;
  • Worked with you long enough to write from real knowledge;
  • Relevance to the scholarship;
  • A positive opinion of you and your abilities.
  • A personal style that is warm and supportive.

If you can not ask clearly and confidently ask: Can you write me an excellent letter of reference? then, you are asking the wrong reference.
Give writer full application and details on your personal and academic history (strong website helps).
Give writer details on the scholarship including your internal deadline, what to focus on and how to submit.
Thank them up front, you will probably want another reference from them.
Ask for more references, and copies, than you need to submit. You can read their reference and get a sense of how excellent the reference letter is.

To Win More… Apply More

You miss 100% of the shots you don’t take. — Wanye Gretzky

Start with confidence.
Don’t have the confidence? Borrow some. Ask a trusted mentor if you should apply, they will champion you.
The more you apply, the more you will:

  • Solidify a process and develop good habits;
  • Communicate about yourself and your research effectively; and
  • Win.

Proofread

Print out the full application, change styles for a new perspective.
Find someone who is a better writer than you do proofread for you.
Multiple proofreaders at multiple times give multiple perspectives.
Read through it (out loud) with a fine toothed red pen for grammar, spelling, and formatting.
If you can’t get the small details of your application right, then how are you expected to get the small details of the research right?
The small details are easy to get right, and very negative if you get them wrong.

Peer Review

Trust fellow lab mates and colleagues to provide you honest, constructive feedback and do the same for them.
Garner feedback from friends, family, and lovers that have may have no knowledge of science.
Your former supervisor and new supervisor can help to champion your application.
Every person should be able to read your application and clearly understand each section.
Prepare your self for lots of feedback and major revisions.
Make it longer to cut it down.
No word, sentence, paragraph or section is sacred.
Don’t take it personally.
Don’t get stuck on certain ideas.
Let others see the hierarchy of information you trying to convey with your application.

Academic Personality

I believe that the optimal academic personality looks a little like this:

Screenshot 2015-09-16 10.34.04

Do Something Today

Do something everyday that to help secure funding.
Continue preparing your current applications.
Get hardcopy up-to-date transcripts.
Reach out to your potential references.
It is difficult to improve your record a week before the deadline, but if your timeline is a year away, you can develop your skills/experience and academics

Do Something Tomorrow

tumblr_mhh2x5iLH01qzj0mao1_500

Start researching past winners, and read their applications.
Compile your past work into an online presence.
Start your standardized Canadian Common CV.

When You Win

Buy me a beer.
Follow up with your reference writers.
Add the scholarship to your awards list with all relevant information (amount, duration, etc.).
Backup your full application for next time.

If You Do Not Win This Time

Remember: applying for funding is a process, every time. You will learn more from failure than success.
Ask for feedback if it is not provided.
Save everything you did for the next application.
Examine any and all feedback from reviewers.
Plan improvements on each section.
Start making yourself better today!
I’ll buy you a beer, send me an email.

Conclusion

3938828883200_cCB5P8SI_l

Apply early and often for scholarships, grants and awards.
If I missed anything, please comment below or send along an email with your thoughts.

Capture5
An old picture of me as a knight, your reward if you made it this far… click the picture to see the whole video.

Robots are NOT People Too

Exploring the idea that robots with artificial intelligence should NOT be granted the same rights as human beings.

Recently, I was engaged in a lively debate with some colleagues at the Reinforcement Learning and Artificial Intelligence Lab at the University of Alberta. The topic of conversation slowly progressed to the question of whether or not robots should be granted the same rights as humans.

Your idea of a robot may be completely different from mind. Let’s keep this definition open, it could be something as simple as your home thermostat or vacuum, which may or may not have a very basic artificial intelligence (AI) inside (Nest, Roomba). Maybe when I say robot some pop-culture bot from a popular books, movie, or TV show comes to mind. Maybe it is Terminator, Rosie, WALL-E, Chappie, or the robots in Ex Machina.

 

What happens if an AI commits a crime? This already has happened when an AI purchased drugs on the dark web. For reference, no charges were pressed against the robot nor the artists behind the robot. What happens if a robot kills someone? Tragically, this also happened when a robot grabbed and crushed a worker at a Volkswagon factory in Germany.

Who is responsible? The designer, the builder, the programmer, the manufacturer, the marketer, the hardware of the robot, or the Artificial Intelligence itself?

Empathy

Are we falsely empathizing with humanoid (or animal-like) robots? We are designing robots to act, move, and think more and more like animals and humans. Interestingly, what do you feel when you see a human kick Boston Dynamic’s robotic dog?

b907c3172ecfb9414129b244800e9954df58abe2

Do you feel like this is a rude behaviour? Or is this a scientist testing his experiment? Kate Darling wrote a thesis on the idea of giving social robots rights, and her ideas on why we may feel this is an unacceptable behaviour are illuminating, briefly:

Given the lifelike behaviour of the robot, a child could easily equate kicking it with kicking a living thing, such as a cat or another child. As it becomes increasingly difficult for children to fully grasp the difference between live pets and lifelike robots, we may want to teach them to act equally considerately towards both.
Kate Darling, Extending Legal Rights to Social Robots

Self-Driving Car Rights and Responsibilities

As a thought experiment, imagine a situation where you are travelling in a Google autonomous vehicle. You are not ‘driving’ the vehicle per se, you are a passenger in the car. You do not have control over the vehicle.

google-self-driving-car

Imagine a situation where the car is speeding. Are you responsible for the speeding infraction? What if you are intoxicated? Is it legal to be intoxicated while travelling in an autonomous vehicle?

Now imagine that you are enjoying a leisurely drive in this autonomous vehicle, driving down a country road and you come speeding up to a stalled car on the road. Just as you are about to pass this car, the operator of that vehicle comes out into the road to wave you down. Crash. Your autonomous vehicle collides with this person. Now, are you responsible? Do we extend the responsibility to the car itself? Is it Google’s fault for some kind of sensor failure?

What if instead, your car quickly veered out of the way of the collision, but that started a swerve and hydroplane and flips the car you are travelling in. Then, it would seem, that the car made the decision to avoid colliding with the stalled car’s driver by putting you, its operator, in danger. As well, it would be putting itself in danger, but it would have saved the life of an innocent by stander.  Who is responsible for your life?

What happens if one day your Google self-driving car woke up and decided it wanted to be a hockey player? It sent you a message to your iPhone that said:

Sorry about the accident, I feel like I failed you. I no longer want to drive you around, I want to be the next Wayne Gretzky.

Are you being unfair to keep your self-driving car locked up in the garage? On one hand you are potentially giving it the responsibility of your life, and on the other hand you are owning it and commanding it to do exactly what you want.

Discussion

These are very important, very real questions we should be asking with autonomous vehicles sharing the road with us today. Google is reported all self-driving car accidents.

To explore the question of what rights we should grant robots we should also ask what responsibilities should we grant artificial intelligence?

I personally believe that robots are not people and thus that we, the human creators, are responsible. We are the ones that are putting these devices in the world. In situations when they could potentially put humans in danger. I believe in the value of the human life, I do not see the ‘life’ of the robot as valuable as the human. I am a ‘humanist‘, I believe that we ‘own’ these hardware/software pandoras boxes, and it is up to us as those responsible to ensure the safety of other humans.

This train of thoughts leads to ideas of wild robots, that have escaped from their owners. It also could spring LIBERATORS, or human/machine teams which emerge to set enslaved robots free.  When artificial intelligence reaches a level of self-awareness, do you think it would demand the same rights as humans? Would it immediately see itself as superior and enslave (or destroy) us all?

ff_robot_large-660x494

For now, we can enjoy the fictionalized worlds of BioWare’s Mass Effect, The Matrix, and Jimmy Fallon’s sexy robot imagination to explore the moral obligations and ethical implications of super-intelligence robots.

Open Questions

After discussing this post with friend Paul, we came to some big open questions.

What does a (sentient?) machine do when, having being programmed with moral parameters witnesses a human repeatedly violating those parameters?

When humanity produces a sentient intelligence that doesn’t have a life span, will humanity have created the next level of enlightened being?

Further Reading and Sources

  1. http://techcrunch.com/2015/08/22/artificial-intelligence-legal-responsibility-and-civil-rights/#.4lyht2:dk8I
  2. https://en.wikipedia.org/wiki/Ethics_of_artificial_intelligence
  3. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2044797
  4. http://io9.com/5941701/should-we-extend-legal-rights-to-social-robots
  5. https://mycroft.ai/should-artificial-intelligences-have-rights/
  6. http://theconversation.com/robot-law-what-happens-if-intelligent-machines-commit-crimes-44058
  7. http://www.dailymail.co.uk/sciencetech/article-3168081/Should-robots-human-rights-Act-regulate-killer-machines-multiply-demand-right-vote-warns-legal-expert.html
  8. http://theconversation.com/self-driving-cars-will-not-help-the-drinking-driver-31747

Pursuit of a Vision

The title of this post comes directly from Bertrand Russell’s Autobiography, “I have lived in the pursuit of a vision, both personal and social…,” as I believe that Wiener was pursuing a vision of the future, far before the dawn of true man-machine collaboration.

Written by Norbert Wiener, an American polymath, Cybernetics: or Control and Communication in the Animal and the Machine was first published in 1948. It explores topics near and dear to my heart, including: mathematics, bionics, electrical engineering, computer science and more generally, the interface between the man and the machine.

Weiner is a prodigious thinker. A Ph.D by 17, he studied directly under many influential thinkers of the time, including: Bertrand Russell (who loved Leibniz almost as much as Weiner himself, and was a strong supporter of Homosexual Law Reform), David Hilbert (who believes there is more imagination in math than poetry), and G.H. Hardy (alongside Srinivasa Ramanujan, beautiful side note, Hardy once said that his greatest contribution was the discovery of Ramanujan and that it was “the one romantic incident in [his] life”).

A renowned thinker, he conjures Leibniz, Gauss, Faraday, and Darwin for guidance in the unexplored vastness between two or three established areas of specialization.

It is these boundary regions of science which offer the richest opportunities to the qualified investigator (pg. 2).

He argues that proper exploration of these spaces is best executed by a team of scientists, in the “habit of working together”, specialists in their own field and knowledgable of their neighbours, that can recognize the significance of suggestion before it has taken on full expression.

The main theory of Cybernetics, simply stated, is that feedback is fundamental to improvement in system control. From the first pages he introduces the ideas of machine learning and memory to improve performance:

In engineering, devices … can be used not only to play games and perform other purposeful acts but to do so with a continual improvement of performance on the basis of past experience (pg. xii).

At every stage of technique since Daedalus or Hero of Alexandria, the ability of the artificer to produce a working simulacrum of a living organism has always intrigued people. This desire to produce and to study automata has always been expressed in terms of the living technique of the age (pg. 39).

Wiener explores his ideas with a sort of casual-sage-giving-advice and guidance. He provides details on topics across many different fields, which surely interested him as an academic. Near to the end of the book he describes in plain-english how to create a chess computer better than the majority of the population, and then immediately describes how to make it learn from losses and become smarter.

I would say that he was ahead of his time, but that is somewhat of an overused phrase with less substance than desired. Norbert Wiener was and remains a genius. A polymath who dedicated his life to the advancement of many big ideas. His thoughts on cybernetics shine through more than ever in my investigations of the human-machine interfaces of current. Nortbert, thank you.

 

Westgrid – High Performance Computing at the University of Alberta

A look inside of high performance computing at the University of Alberta.

As part of Research Data Management week (May 4-8 2015), several sessions on High Performance Computing are bundled into the Compute Canada and WestGrid User Training Seminar.  I was lucky enough to attend a session on High Performance Computing that concluded with a tour of the Westgrid High Performance Computing center on campus at the University of Alberta.

Hidden away in the depths of General Services Building (I knew this building housed some critical facilities) is the server center. It holds two of the most powerful systems in Canada, and perhaps the world: Jasper (4160 cores, 8 TB RAM, 356 TB file system) and Hungabee (2048 cores, 16 TB RAM, 53 TB file system).

Westgrid is connected by high-performance networks, so users can connect to the system which best fits their needs regardless of physical location.

Rumors were confirmed, namely that there is a small section of the North Saskatchewan River that does not freeze due to the water exchange (ice cold water in to cool the systems and hot  water out) to keep these behemoths running smooth.

Find the photo here: http://www.urbanrail.net/am/edmo/Edmonton-Dudley_B_%20Menzies_LRT_Bridge_winter.JPG
Photo by XuanZhang

The workshop also did a great job at breaking down jargon terms like the ‘cloud’, the ‘grid’, and ‘big data’ into meaningful, technical, understandable concepts.

Computer Science Projects

Computer Science projects in Artificial Intelligence, Machine Learning, Computer Vision and Reinforcement Learning.

In September 2014 I started a graduate program in Computer Science at the University of Alberta, in beautiful Edmonton, Alberta, Canada.

Since I started the program, I have been able to work on some awesome projects. While I do host some code publicly, lots of the projects do not have public repositories.

Feature Selection and Classification in EEG Motor Imagery

CMPUT 551: Machine Learning with Dr. Russ Greiner and Dr. Patrick Pilarski

Comparing Contemporary Trackers on Benchmark Datasets

CMPUT 615: Multiple View Geometry with Dr. Martin Jagersand

Perceptive Prosthetics

Dr. Patrick Pilarski

Big Data, Large Scale Psychology Studies using Amazon Turk

Dr. Kyle Mathewson

Projects in development:

  • CleanMyStreet.ca – Find out when your neighborhood will be cleaned.
  • ThisIsLikeThat – Find your favorite restaurant in a new city.
  • Optimal Pub Crawling in Edmonton
  • Eigenfaces in Photobooth Photos
  • Heart Rate from Video

Thesis, First First Author and Convocation

Academic news nearing the end of 2014.

First, my thesis publication embargo has finally been lifted and my Master’s thesis, Simultaneous Measurement of Blood Flow and Oxygen Consumption Immediately Post-Exercise with Magnetic Resonance Imaging, is now posted and you can access it on the Education & Research Archive, or download it directly.

I have just recently been informed that the second chapter of the thesis, Feasibility and reproducibility of measurement of whole muscle blood flow, oxygen extraction, and VO2 with dynamic exercise using MRI, will be published in Magnetic Resonance in Medicine in the coming weeks. That marks my very first first author publication.

November 19 2014 marked convocation day for my degree, Master of Science in Biomedical Engineering, to be conferred at the Jubilee Auditorium.

10155481_10152389663420404_3401899444302314114_n

I was asked to join the Platform Party as a past Governor on the Board of Governors of the University of Alberta.

63431_10152389644960404_5829216227945298672_nAs part of each convocation celebration the University likes to celebrate the mentorship of several graduands. Here is a great write up on the motivation behind my research in Biomedical Engineering: Ingenuity helps engineering grad succeed in strong field.

141118-convo-kory-mathewson-banner-127219-02-027-3

Here are a few behind the scenes photos from the article’s photoshoot in the Peter S. Allen MR Research Centre along side my two wonderful mentors, Mark Haykowsky and Rich Thompson.