P.D. Magnus

Two links about AI

There are some articles that I read and think I ought to blog about that. Then I realize that I basically have. So this is basically a link dump kind of post.

Link #1: Geoffrey Hinton cautions that deep learning is not especially deep

I’ve written some posts about the glitzy fad for “deep learning”. It has the same strengths and weaknesses it had when it traveled under the less-shiny banner of “back-propagation neural networks”.

Link #2: Efforts to understand the bias inherent in algorithms

Procedures that are superficially objective can encode bias. I don’t have anything deep to say here, but I’ve blogged about it before.

The slipperiness of significant publication

I’ve written before about trying to establish the significance of a philosophy publication, which mostly becomes an issue at tenure and promotion time. In an earlier post, I argued that citation counts are mostly rubbish for philosophy. Still, there are pressures to provide numerical measures.

In support of a recent tenure case, our department gave the acceptance rates at journals as evidence of their significance.¹ The danger in using such a measure is that, if widely adopted, it would quickly become uninformative.² If scholars get more credit for publishing in journals with low acceptance rates, they will preferentially start to submit to those. The journals will get more submissions but still accept the same number. This will, in turn, drive down the acceptance rates of those journals.

I recently learned that there is a name for this general phenomenon, Goodhart’s Law: “When a measure becomes a target, it ceases to be a good measure.”

This shifts it from being this hunch I have to being a well-established phenomenon. I can now invoke it with greater conversational gravitas. Continue reading “The slipperiness of significant publication”

Nothing leads where I expect

When I started News For Wombats, I expected to be writing slightly different stuff than I did at Footnotes on Epicycles (my old blog).

I thought I’d post about politics, but I haven’t.

The present scene is alternately too depressing and too crazy. One example of a post that almost happened: I was struck, in the aftermath of Charlottesville, that the New York Times could still whinge about Trump’s “moral standing as president.”³ I mulled it over a bit before posting, Trump doubled down on deferring to racism, and then the Time’s understatement seemed hardly worth mentioning.

I thought I’d write more capsule reviews, but I haven’t.

For example, the Netflix show Iron Fist is almost unwatchable. I ultimately soldiered through so that I could pick up the continuity of the Defenders. The biggest problem with Iron Fist is that the action scenes are terrible. I don’t just mean that the martial arts are technically poor (which they are) but also that they fail narratively. I am fine with fight scenes in old Star Trek episodes: Kirk takes out aliens with hammy rabbit punches, but it comes across that he’s supposed to be pretty good at punching. Silly fight choreography aside, the narrative is clear. In Iron Fist, Mister Fist is thrown around by even mook-level bad guys. The show doesn’t let me the viewer know that he’s supposed to actually be good at this.⁴

Hey, now I have written about those things. How about that?⁵ Continue reading “Nothing leads where I expect”

Tweets point nowhere

Mark Simonson’s blog got me thinking about information technology and the original aspirations of hypertext. Simonson laments that current technology is too much driven by concepts taken from print media. Part of the problem is the lack of a clearly defined alternative. Ted Nelson, who coined the word “hypertext”, had a vision of multiple texts floating on-screen with lines connecting points in one to points in another. I don’t see how that wouldn’t end up like items on a cork board linked by lengths of yarn, the idiom for madness from A Beautiful Mind which has become Hollywood shorthand for crazy conspiracy theories.

Old school blogging actually seems like a pretty good realization of hypertext. Good blog post take a while to write because you’ve got to provide pointers so that someone who hasn’t got context or who is curious can follow up. Someone who wants even more can search on key terms.

All of this crystallized for me what I don’t like about Twitter. In order to cut a thought down to Tweet length, people leave out context. What are they enraged about? What’s the thrust that drew their clever riposte? I can’t always tell.

Sometimes thoughts that won’t fit into a single tweet are written as a stream, possibly with numbered entries 1/9, 2/9,… I see entry 4 of 9 because someone reweeted it, and it’s a serious investment of effort just to view the original series in order. Even then, I can’t always suss out the context.

Twitter, in short, is hypotext. It eschews the links of hypertext but also the context you’d expect from a letter or newspaper article.

Part of the shift is that many people go on-line primarily with phones or tablets, appliances that are great for scrolling and clicking but bad for following multiple threads. Twitter and Facebook turn our feeds into one-dimensional things. We can scroll through, liking and reposting as we go. But reposting just drops another log somewhere into the flume.

Reader query, re: anagrams

Based on your own sense of how words work, pick one of the following:

Every word is an anagram of itself.
Some but not all words are anagrams of themselves.
No word is an anagram of itself.

There’s a principled case to be made for every answer. Cristyn and I hashed it out over goat cheese last night, but I won’t tell you the considerations we mustered on various sides or what we concluded. I’m curious about what you think.

Recruitment bleg

My department has had a precipitous drop in both applications and admissions for our PhD program. We have funded positions which we haven’t been able to fill with new-to-program students. These are a pretty good deal and involve no teaching responsibilities for the first year.

So I have two requests:

First, how do you think we should go about trying to make people who might be interested aware of the opportunity? Some programs mail out glossy posters touting their graduate programs. I suspect those mostly get thrown away, however, and I’m not sure that people these days look to posters for information.

Second, if you are in the position of mentioning our program to someone who might be suited for it, then please do so. Obviously this isn’t everybody: Students who can get funded slots at top-ranked programs should take them, some students should be talked out of going to grad school, and so on. But there are students who would get a lot out of coming here. We have strong faculty in a number of areas. Our graduates have a decent record of getting permanent and even tenure-track jobs. Many have gotten jobs at community colleges, the sort of jobs that aren’t even on the radar for graduates from top-ranked programs.

If you have ideas or questions, please comment.

Paean for the open internet

Today is a day of action in which technology companies plump for net neutrality. Action seems to mean talking on the internet.

In filling out a petition-thing, I wrote this:

There’s a form letter I could have cut-and-pasted here, but this is important enough to write my own words. As a user of the internet, I want to be able to access the content which I decide matters. I want it to come at the same speed other content would come at, rather than having it be faster or slower based on whether someone who owns that content has decided to pay more for access to me. If they get control over accessability and relative speed, then I’m not a consumer anymore but instead I’m the product that the service provider sells to their customers. That’s why net neutrality matters.

It seemed worth posting here, too.

Hi-phi, podcasts, mash-ups, and covers

There is a short review of Hi-Phi Nation in today’s Guardian, topped by a nice picture of host Barry Lam.

Cristyn and I were in Poughkeepsie last week to talk with Lam about cover songs. He plans to do a show about musical covers. He called us because he had read the paper we wrote with Christy Mag Uidhir and we’re local.

While there, we also took in the Walkway over the Hudson.

Continue reading “Hi-phi, podcasts, mash-ups, and covers”

The puzzle of virtual art

If I recall correctly, Arthur Danto considers the example of an art deco bronze cat which is chained to the railing on a landing in the student center. This might, he says, be a recent artwork composed of a readymade bronze cat, a chain, and a pad lock. Perhaps a commentary on domestication. Nevertheless, the artwork is just the cat itself. Someone chained it to the railing when it was installed in the student center, so that nobody would walk off with it.

Similar examples arise in gallery contexts: It may be unclear whether two objects are one composition or two works simply installed near to one another. A toolbox might be left behind by a worker and a utility box may just be hardware that’s part of the gallery, or these things might be works or parts of works in the show.

I have seen these puzzles mostly raised in the context of ontology, to pose the question of what constitutes a work of art. Yet, it seems to me, there is an important issue in art appreciation.

Suppose, to return to the original example, that one finds it rewarding to consider the chain and cat together. They are visually interesting, let’s say, and one is lead to reflect in interesting ways on domestic animals and public spaces. It’s rewarding in just the kinds of ways that reflecting on an artworks is. Suppose furthermore that considering just the cat on its own is not especially rewarding. It is not an inspired sculpture, and it is of limited historical interest. One gets more out of considering the chained cat as if it were art than in considering the actual artwork.

What should we think about this? Several possibilities:

First, one might say that considering non-art as if it were art is perverse and confused. It is either a conceptual or moral failure. I find this implausible.

Second, one might say that the rewards of reflecting on the chained cat show that it must be art after all. We are supposing that nobody explicitly meant it to be art. We could say that the staff member who chained the sculpture had an unconscious intention to make art. Or we could say that one makes it art just by engaging in the rewarding as-if art experience of it.

Third, one might give up the idea that artworks are especially important to art appreciation. What matters are objects which can be encountered and interpreted in as-if art ways, regardless of whether they are actually art.

AI ai ai

I recently commented on the fact that machine learning with neural networks now regularly gets called “AI”. I find the locution perplexing, because these machine learning problems have success conditions set up by engineers who defined the inputs and outputs.

Here is another headline which doubles down on the locution, discussing AIs creating AIs. Yet having a neural network solve an optimization problem is still machine learning in a constrained and specified problem space, even if it’s optimizing the structure of other neural networks.

Brave new age of robot overlords this ain’t.

Continue reading “AI ai ai”