May 20, 2013

Google I/O 2013 Trip Report

Since a handful of people have asked me about Google I/O, and it isn't a conference that a lot of academics go to, I thought I'd share my experience with this group.

First off - why did I go?  The answer is because I didn't go to CHI so I wasn't worn out, and because Google asked.  They offered me a $300 ticket (compared to the normal $900 which is also very hard to get).  The primary reason I wanted to go was to experience the silicon valley big company event. I was interested of course to learn about Google's view of modern web technology, but mostly I just wanted to feel what this kind of event was like. (And full disclosure, I have research funding from Google.)

So, first let me comment on that feeling.  It was pretty much what I expected - and totally surreal. It was VERY well organized with hundreds of Googlers and staff helping everywhere.  Academics were treated specially (with a short special line, a special floor at an evening party, an invite to Google's SF offices for a tour, etc.). And it was seriously geeky. I was a distinct minority - caucasian and old.  There was very little gray hair and very few women. And a lot of Asians of all kinds. The keynote was a lot like a rock concert. Super crowded - intense spirit. Thousands of people holding up their phones to take pictures of the screens showing fluid blobby animations to rock music before the start. And every one yelling a count down to the start (illustrated on the displays.)  I almost laughed at that part - I guess you can tell my age. I enjoyed it, but just couldn't get *that* worked up about it :) The 3 hour keynote (and all the talks) was very well orchestrated and presented. I really enjoyed hearing Larry Page give his talk. Unlike the other highly scripted talks, his was clearly quite heartfelt. He was doing his best to step out of his demanding role and relate to his techy audience, and I think he succeeded. (Of course, cynics could argue that he was just very good at scripting the feeling of non-scriptedness...)  He largely talked about the importance of being innovative and doing good - and stop the cat fighting that tech companies do. And then without any warning, he took 30 minutes of questions and there was a pretty interesting discussion - probably my favorite part of the whole conference.

There was, of course, excellent free food (breakfast and lunch). Free coffee. Free snack area. And a free Chromebook Pixel (retail $1,450) - meaning they gave away something like $10m worth of them. This event was not a money maker! At least not directly. They of course spent 3 days exciting 6,000 developers about using Google tools to help pursue Google's vision. You may have heard that Google is making their interfaces simpler - removing more stuff, predicting more stuff, and using more of your personal data. Some people will no doubt find that very scary and objectionable. But clearly lots of people (including, I admit, me) find it an acceptable trade-off.  And their demos were astoundingly good. For any of you that remember Microsoft's painfully embarrassing voice recognition demos a few years back - Google ACED them. There was 100% recognition rate of live demos of the speaker talking into an actual phone, speaking fairly complex natural language at regular rates in a very noisy room of 6,000 people.  And if you remember "Put That There" from 1979, then it was all the more impressive when they had good pronoun connectivity back to previous utterances.  Expect to see a lot more voice recognition built into every Google product.  (And no more clicking - soon you'll be able to just say "ok google" and start talking when most google products are open.)

Glass was a big deal (but hardly mentioned in the keynote).  Hundreds of people were wearing them. There was a place to try them out (which I did). And a number of talks on Glass (videos available for many of them.) Personally, I find Glass intriguing - but in its current form I think it is unlikely to be used outside of niche areas and geeks (beyond a startup phase). I predict they will be like bluetooth headsets. They will probably be everywhere next year until people realize that it is too annoying to wear something all the time (and charge them, etc) for the infrequent times when the value add over pulling out your phone is worth it (again, except for the geeks and niche uses). OTOH, if the cost was much less (i.e., it was built invisibly into regular glasses, had inductive charging, etc.) then that value calculus could be very different.

And then there were 2.5 days of talks, which I got somewhat bored of by the second day. Two of the talks I most wanted to see were overfull and I couldn't get in. It was very crowded and difficult to move around between sessions. And many of the others were just too deep about topics I wasn't that interested in (i.e., APIs for every Google product). It would be greatly rich for people building web startups using all of those tools - but academic research tends to be much more selective. And when there is something I actually want to use, it is easy enough to learn it with public resources.

So, bottom line, I am glad I went for the overall experience. But I'm not sure that I would recommend it to most academics (especially since most of the talks are available online.) But it was fun. And Google probably succeeded in their biggest goal which is that I left feeling more positive about Google than when I arrived.

April 29, 2013

Why Lecture Isn't Sufficient

It takes effort to learn. A common adage is that the best way to learn something is to teach it. So, it shouldn't be surprising that simply listening to someone lecture isn't the best way to learn. This is nothing new, but I've been reading more and more that show that alternative "active learning" approaches are demonstrably better. And this works in a wide variety of settings at surprisingly large scale.  I've collected here some particularly interesting references (in reverse chronological order):

Beth Simon, Julian Parris, and Jaime Spacco. 2013. How we teach impacts student learning: peer instruction vs. lecture in CS0. In Proceeding of the 44th ACM technical symposium on Computer science education (SIGCSE '13). ACM, New York, NY, USA, 41-46.
=> Replaced lecture with peer instruction (PI) in a non-majors CS0 course (i.e., the one before a traditional freshman majors class) resulted in approximately a half letter grade improvement (N=90). The PI class was assigned reading before class that was assessed with a short quiz at the beginning of each class. Much of the lecture was replaced with a sequence of: 1) individual clicker responses; 2) discussion in assigned groups; and 3) individual clicker responses again; 4) whole-class discussion.

Leo Porter, Cynthia Bailey Lee, Beth Simon, and Daniel Zingaro. 2011. Peer instruction: do students really learn from peer discussion in computing? In Proceedings of the seventh international workshop on Computing education research (ICER '11). ACM, New York, NY, USA, 45-52.
=> Replaced lecture with peer instruction in two upper-division majors courses (Intro to Computer Architecture, N=51; and Intro to Theory of Computation, N=45). PI implemented similarly as first paper (above). Findings that 85-89% of "potential learners" benefit from peer discussion. 90% of students broadly agree that "I recommend that other instructors use our approach ... in their courses."

Louis Deslauriers, Ellen Schelew, and Carl Wieman. 2011 Improved Learning in a Large-Enrollment Physics Class. In Science 13 May 2011, 332 (6031), pp. 862-864.
=> In two Quantum Mechanics physics courses (N=~270 each), flipping one week of the course (replacing lecture with a variety of active learning activities) resulted in increased student attendance, higher engagement, and improvement in learning from average test scores of 41 to 74. The active learning classes included "pre-class reading assignments, pre-class reading quizzes, in-class clicker questions with student-student discussion, small-group active learning tasks, and targeted in-class instructor feedback. ... The small-group tasks were questions that required a written response. Students worked in the same groups but submitted individual answers at the end of each class for participation credit."

Jason A. Day and James D. Foley. 2006 Evaluating a Web Lecture Intervention in a Human-Computer Interaction Course. In IEEE Transactions on Education, Vol. 49, No. 4, 420-431.
=> Flipping an HCI course resulted in grades that were just about a full letter grade higher in comparison to a control condition. The control reduced the amount of lecture time by the estimated time students watched material outside of class. The lectures were replaced with web-based video lectures including PowerPoint slides and video of the instructor - much like today's MOOCs. Students had short quizzes on the web lectures before class. These "Lecture Homework Assignments" (LHWs)  were "synthesis-type questions that require students to discern and elaborate on concepts covered in the lecture" - "not simply verification or summary-type questions". After the LHWs, class time included "project-related group presentations, small breakout group discussions and presentations, redesign sessions, design critiques, design reviews with HCI experts, role-playing activities, discussions with local HCI practitioners, and others."

April 17, 2013

Special Advisor to the Provost on Technology and Educational Transformation

I'm excited to announce the new role I've just taken on at UMD.  Here is the letter I just sent to our faculty describing it.  I've already received about 2 dozen thoughtful replies, and am really looking forward to meeting the people innovating on campus!

Dear Colleagues,

It shouldn't be a surprise that our campus is looking closely at the use of technology in education, and our faculty are actively experimenting with different approaches. Recently, there has been a confluence of pedagogical interest and technological advancement, which has made the application of technology to education a strategic interest at UMD and at universities around the world. There are MOOCs (such as Coursera, Udacity and EdX), online classes, and any number of hybrid approaches. "Blended" educational environments combine face-to-face and electronic classrooms, and "flipping" the classroom (lectures online and face-to-face classes for more active, engaged learning) are just some of the strategies that instructors are pursuing.

The opportunities (and risks) here are significant enough that the Provost has asked me to help her develop a strategy to best infuse technology into education throughout campus. Thus, starting immediately, I will be taking on the role of Special Advisor to the Provost on Technology and Educational Transformation, through August 2014. I hope to consider a wide range of activities while keeping the focus on our core mission of improving the quality of the education we offer to our students on campus. But it goes further than that. The opportunities for innovation, study, and publishable research in this space are tremendous. And the Provost has made it clear that pursuing research-based understanding of these issues is crucial.

As a computer scientist working in the area of Human-Computer Interaction, I have been involved in developing and applying technology to support education. From the International Children's Digital Library to SearchParty, I have endeavored to create new ways that technology can be used to motivate students and increase their power to learn.

As the campus considers these issues more broadly, I welcome your perspective. If you have exciting new ideas, or concerns about technology in education, let me know. As I said, the opportunities are tremendous - but it is up to us to invent the future.

Looking forward to working with you,

  - Ben