the (Frustrated) Over-Analyst: MoAR #15: thoughts on AI, controversy, and a new find ... among other things.

between attempting to file my american taxes and traveling, there wasn't much time to roundup the usual slew of articles. here's what i found, though:

apparently it's not called A.I. anymore ..

.. but AGI, short for artificial general intelligence. all in all, i like the article in that it does a good job of explaining the background of computing, deconstructs popular notions of machine intelligence, and discusses (in broad strokes) some notions of universality.

"Why? I call the core functionality in question creativity: the ability to produce new explanations. For example, suppose that you want someone to write you a computer program to convert temperature measurements from Centigrade to Fahrenheit. Even the Difference Engine could have been programmed to do that. A universal computer like the Analytical Engine could achieve it in many more ways.
..
Now imagine that you require a program with a more ambitious functionality: to address some outstanding problem in theoretical physics — say the nature of Dark Matter — with a new explanation that is plausible and rigorous enough to meet the criteria for publication in an academic journal."

~ from "The very laws of physics imply that artificial intelligence must be possible" @aeonmagazine

it's an interesting spin on the turing test but whether there is a specific focus on creativity, a phenomenon that we all know exists but is very difficult (if not impossible) to measure.

imagine: a creative computer. i guess the best science is the kind that inspires with great challenge.

"The upshot is that, unlike any functionality that has ever been programmed to date, this one can be achieved neither by a specification nor a test of the outputs. What is needed is nothing less than a breakthrough in philosophy, a new epistemological theory that explains how brains create explanatory knowledge and hence defines, in principle, without ever running them as programs, which algorithms possess that functionality and which do not."

controversy.

you may have seen this headline around in the last week or two, and maybe even this excerpt ..

"Over the years, I have co-written many papers with mathematicians and statisticians, so I can offer the following principle with confidence. Call it Wilson's Principle No. 1: It is far easier for scientists to acquire needed collaboration from mathematicians and statisticians than it is for mathematicians and statisticians to find scientists able to make use of their equations."

~ from "Great Scientist ≠ Good at Math " @wsj

(as you can imagine, there have been many rebuttals to wilsο's essay, including this one.)

the thing is: i actually agree with this "principle" if only because pure mathematicians generally work with notions that, in their present form, aren't physically intuitive. the reactions and feedback loops in chemistry and biology may exhibit incredible complexity, but ultimately they are about particles and cells that can be measured and observed.

however, try explaining non-archιmedean norms on the real line to someone who doesn't care about number theory, and you might have a problem.

there is another point i want to make, though: when i read that excerpt, i take it to mean that scientists need us, but we don't actually need them.

so this fall i might be joining an endangered species.

everyone is aware that times are tough. some can even prove it .. at least empirically.

"And with stretched budgets and public pressure to keep costs down, many colleges and universities are cutting back on tenure and tenure-track jobs. According to the report, such positions now make up only 24 percent of the academic work force, with the bulk of the teaching load shifted to adjuncts, part-timers, graduate students and full-time professors not on the tenure track."

~ from "Gap Widens for Faculty at Colleges, Report Finds" @nyt

"This doesn't actually mean that there are fewer full-time professors today than four-decades ago. College faculties have grown considerably over the years, and as the AAUP notes, the ranks of the tenured and tenure-track professoriate are up 26 percent since 1975. Part-time appointments, however, have exploded by 300 percent."

~ from "The Ever-Shrinking Role of Tenured College Professors" @theatlantic

i guess i confirmed my own confirmation bias. since i essentially encounter only tenure(d / track) faculty and postdocs, the samples and sizes can easily skew the perspective that "everyone eventually gets a job .."

find of the week: fastco

fastco is another one of these techie/startu p websites that (think they can) change the world for the better, through quantitativity and effort. the discussion, however, can get interesting without going too technical.

for instance, this excerpt is about nutrition and how to measure your data .. not so that you can mine it, but so that the reader can interpret it more viscerally:

"People who viewed the menu without nutritional information ordered a meal totaling 1,020 calories, on average, significantly more than the average 826 calories ordered by those who viewed menus that included information about walking-distance," writes Scientific American. People who saw the menu with walking-distance info also ordered less than people who just saw calorie info."

~ from "We Should Measure Our Food In Exercise, Not In Calories" @fastco

as for the next one, it jarred my thinking a bit and it still puzzles me.

""Flipped classrooms" rethink traditional models of classwork and homework. Teachers are valuable and need to be better utilized. In a “flipped classroom,” passive activities, like lectures, are reserved for homework, while in-class time is used for collaborative and personal interactions between teachers and students. Teachers can post their own lectures online and direct students to other online resources, such as those provided by Khan Academy, which offers more than 2800 educational videos covering a multitude of disciplines. Students can pause, rewind, and re-watch as needed. Any questions can be noted and addressed the following day. "

~ from "5 Disruptive Education Trends That Address American Inequality" @fastco

it sounds like a good idea, but how stable is it in practice? in other words, suppose a student misses a lecture once, the night before, and shows up unprepared: can that student get anything out of class the next day?

also, why do i have the feeling that "flipping the classroom" amounts to posting the lecture as a video online, and shifting the recitation/problem session to the front-&-center? (there is a point, in that most lectures are not interactive, though.)

so like i said: it puzzles me. i might even give one or two lectures a try with a topic that is rather computational in nature, next semester.

lastly, a few more thousand words.

first, some interesting algebra: