Argument for Ambiguity

I got directed to a recent piece about tolerance for ambiguity as a job requirement and a skill education should help develop through this quote from a responding blog post: “To the extent that we can provide assignments and experiences in and among classes that give students the experience of getting a little lost and finding their way back, we may be able to build some of that tolerance for ambiguity in the kind of settings Selingo discusses.”

While the original article focuses more on the idea of a “growth mind-set” and encouraging students to think of perseverance rather than innate intelligence as their most valuable asset, from a higher education perspective, I find the reflections about the value of introducing ambiguity into assignments more compelling. Another quote that echoes what I see in my students when presenting them with open-ended, and thus ambiguous, assignments: “In thinking about my own tolerance for ambiguity, I wouldn’t call it high or low. It varies, and I think the major independent variable is my own feeling of competence in the situation. When I feel like I can handle whatever the situation is likely to throw at me, ambiguity isn’t a problem. When I’m utterly lost, ambiguity can feel threatening. The key issue isn’t so much ambiguity or the lack thereof, but its possible outcome and my own sense of vulnerability.”

I see precisely this tension in students every semester, particularly those that are new to our courses and the expectation that they take responsibility for exploring options and refining the scope of a problem for themselves that is common through most of them. It’s a tricky balancing act to present just enough uncertainty in assignments that they get to have this valuable experience, but not so much that the feeling of vulnerability blocks their openness to exploring. This framing of why the ambiguity is intentional, and its role as an employment skill, is an interesting angle on explaining the assignment design to students.

On a bit of a tangent, and returning to the original article, there is one sentence that jumped out to me as odd: “As artificial intelligence increasingly makes many jobs obsolete, success in the future will belong to those able to tolerate ambiguity in their work.” I suspect the point here is that tolerance for ambiguity is one of the higher-level problem-solving skills that are hard to automate out of the work force. But from an artificial intelligence standpoint, this statement is odd because the gap between artificial intelligence versus simply computer automation frequently comes from AI being able to tolerate ambiguity and still function. This doesn’t invalidate the larger point – if ability to function outside strict parameters is one of our tests for successful artificial intelligence, no surprise that employers would like the same characteristic in their intelligent human employees. But on a technical level, this statement jumped out at me as missing some of what is exciting in AI work.

Exercising my writing muscle

I was flipping through Spolsky’s Joel on Software today and, perhaps because I spent the morning working with our college-wide curriculum and some of our documentation of its outcomes, this passage jumped out at me:

So why don’t people write specs? It’s not to save time, because it doesn’t, and I think most coders recognize this. […] I think it’s because so many people don’t like to write. Staring at a blank screen is horribly frustrating. Personally, I overcame my fear of writing by taking a class in college that required a 3-5 page essay once a week. Writing is a muscle. The more you write, the more you’ll be able to write. If you need to write specs and you can’t, start a journal, create a weblog, take a creative writing class, or just write a nice letter to every relative and college roommate you’ve blown off for the last 4 years. Anything that involves putting words down on paper will improve your spec writing skills. [from Painless Functional Specifications Part 1]

Yes, yes, yes! There’s no indication of the content of the class Spolsky is referring to, but it wouldn’t be surprising if it were a humanities course that he took to meet a gen ed requirement. This isn’t the only value of technical students taking courses outside their major, but the likely increase in practice writing is a great one. And I love this example of someone reflecting back on the benefits of a course that probably wasn’t the motivation for signing up for the course, and perhaps wasn’t even recognized at the time.

It also got me thinking about a related question: why don’t people read specs? I distribute programming assignments that resemble specs, and my lab tutors have learned that one of the first steps to helping a struggling student is to get them to go back and actually read what is in the specification. If you don’t like to read, digesting a written description of what you’re being asked to do will be painful. But you can similarly say that reading (particularly reading carefully) is a muscle that requires exercise. So, if reading well requires practicing reading well, perhaps we should be cautious of trends towards assigning less reading and more short excerpts, web links, and videos. In addition to conveying the content within the text, it might well help students do better on assignments that on the surface don’t have anything to do with reading.

Leaving time for focus

This quote from a recent Chronicle article Infantilized by Academe struck me, particularly with the chaos of the end of the academic year:

Our students are often more distracted than we are, and so inured to distraction that they are unlikely to notice it. As other commentators have argued, the process of gaining admission to selective American colleges now requires presenting an array of accomplishments so vast and varied that any reflection that might accompany them is purely incidental.

This thought resonates with recent conversations I’ve been having with students and colleagues about the amount that students try to take on, and the difficulty many students have in recognizing the real cost of doing more. Yes, you can take on more courses and activities, but you will sacrifice the depth of attention each one can receive.

It creates an advising problem for me. By nature, I encourage my students to challenge themselves. Take the hard course they are interested in. Take on research projects. But I’m also finding myself trying to figure out how to advise moderation without advising complacency. Don’t sign up for three upper-level, project-based courses in the same semester – pick the one (or maybe two) that you care about, and make sure you get everything you can out of those courses. Don’t try to complete three or four programs – particularly if you’ll find yourself completing multiple capstones in the same semester and not able to fully dedicate yourself to any of them.

Or, at least, make those choices with your eyes open about what you’ll be sacrificing by going after quantity and decide that it’s a sacrifice you’re comfortable making.

Rescue Robots in the News

This semester my intro programming students are doing a very scaled down model of how search-and-rescue robots might very stupidly explore a space while trying to keep themselves from clumping up with each other. It’s a first programming course for most of them, so have I mentioned that these simulated robots are very stupid.

However, since I’ve been playing around with their project, I seem to be seeing interesting content about search and rescue robots cropping up all over the place:

Last week (on April 23rd), there was a great NASA JPL livestream of a talk on Rescue Robots focusing in particular on RoboSimian.

A prototype of a new robot assistant to guide firefighters through buildings to aid with search and rescue was demonstrated.

Sadly, a shape transforming robot got stuck inside a Fukushima nuclear reactor while trying to investigate the state of the plant to help with decommissioning.

Where’s my plow?

The snow situation isn’t as bad here in Western PA as it is on much of the east coast, but while waiting for things to lighten up enough for me to go out and shovel, I’ve been playing around with Pittsburgh’s new snow plow tracker. The system itself is only live while snow is falling – access it through the button on the right.

I like the use of the “multiple vehicle” icon to keep things legible when zoomed out. It took me a bit of playing around to realize that if you adjust the “history display” slider at the bottom of the screen, you can see the routes the plows took from the target time until the current time. Which, in effect, lets you figure out which roads have been plowed in the last, say, three hours. I’d love to see an overlay of this with the Google traffic information since it’s based on a Google map but maybe that can come in version two.

It’s a testiment to how fun the system is that I’ve lost so much time playing with it even though it doesn’t actual cover any roads I’ll plausibly drive on in a snowstorm. Very interesting to see which roads never get plowed at all.

Counting down

I am a crazy fan of advent calendars. In addition to my physical calendar of ornaments, I’ve got a collection of online calendars I’m “opening” each day as well. Here are my favorites I found this year:

Saveur Cookie Advent Calendar: A new cookie recipe each day – check out day six’s Alfajores

Erik Svedäng’s Advent Calendar: Fun little widgets to watch and, in some cases, interact with

Advent of Indies: Each day another indie game is promoted alongside a freebie to enjoy (some available only on the day the door opens)

LEGO Star Wars Game Advent Calendar: play through different levels unlocked each day to collect pieces

The Economist Daily Chart Advent Calendar: An infographic roundup from the year, with a new one scheduled for release on the 25th

Lorem ipsum ipsum ipsum lorem

“While Google translate may be incorrect in the translations of these words, it’s puzzling why these words would be translated to things such as ‘China,’ ‘NATO,’ and ‘The Free Internet,’”

There is so much to love in this exploration of what happens when you feed lorem ipsum text into Google Translate from Krebs on Security (or, at least what used to happen). Automatic translation algorithms, data sparsity problems, covert information channels… A bizarre, must-read article.

Patchwriting and attribution

If I were teaching a writing skills course this fall, I would be tempted to assign this Language Log post about another recent plagiarism accusation just because of the side-by-side comparison of language and discussion of “patchwriting”. It would probably surprise some students to see the degree of difference between the compared text, and that this is a concern even though the text in question is cited elsewhere, just not for some very specific phrases. Also interesting is the analysis of the older text for whether it too used and attributed patchwriting appropriately – we’re clearly more easily able to spot these things now with digital texts.

Free Service Botnets

How Hackers Hid a Money-Mining Botnet in the Clouds of Amazon and Others: a couple of security researchers build a botnet out of free accounts, potentially legally they claim, rather than from hijacked computers. They proof of concept tested Litecoin mining, suggesting they could have brought in $1750/week with their constructed botnet if left running.

While the article cites Amazon and Google’s services as examples, the following suggests an alternate source for these vulnerable accounts:

Choosing among the easy two-thirds, they targeted about 15 services that let them sign up for a free account or a free trial. The researchers won’t name those vulnerable services, to avoid helping malicious hackers follow in their footsteps. “A lot of these companies are startups trying to get as many users as quickly as possible,” says Salazar. “They’re not really thinking about defending against these kinds of attacks.”

A brief mention late in the article about companies (not Amazon or Google) turning off services or shutting down because of this type of malicious use suggests this may be a real barrier to entry into the market for cloud computing.

Buy your donuts with cash

I read this story wanting to understand if the data mining they’re doing is really appropriate for making individualized statements in the way they are claiming when they suggest that hospitals will get risk assessments based on patient shopping data through credit cards, store cards, etc. Will receiving doctors get sufficient training in the ways in which these predictions are like and unlike the predictions that medical tests make about health risks?

Additionally, I read through the list of hypothesized triggers for heath risks and they seem to bank on the idea that everything I’m purchasing is for myself. Just in the first paragraph the article suggets issues if “ou’ve let your gym membership lapse, made a habit of picking up candy bars at the check-out counter or begin shopping at plus-sized stores” which could nicely match my patterns this year of regularly picking up candy for the lab and buying some clothes for an elderly relative who can’t get out to stores as easily anymore. The majority of the time I buy donuts, I do not actually eat any of the donuts – and I have colleagues whose donut-eating patterns may more closely match my shopping trends than theirs.

And then you get this statement: “While the hospital can share a patient’s risk assessment with their doctor, they aren’t allowed to disclose details of the data, such as specific transactions by an individual, under the hospital’s contract with its data provider.”

I’m not sure if that’s good — hooray for not sharing personal details! — or worse — so the computer says I’m at risk but we can’t sit down and talk about whether the patterns it’s identifying are real. And, going back to my opening question – what sorts of algorithms are being used and given that, what sorts of conclusions are even valid to draw.