The Arizona Desert Lamp

White Paper: School of Information Sciences

Posted in UA Transformation Plan by Connor Mendenhall on 20 October 2008

Leave it to the Computer Science department to come up with a true techie idea—a “virtual school.” From their white paper proposal:

Our plan for transforming the Computer Science department is inextricably linked with our proposal for a School of Information Sciences, Technology and Arts (SISTA). SISTA will be a virtual school.  Most participants will remain in their own units, although over time joint hiring may create faculty lines in SISTA.  Computer Science faculty will become SISTA faculty, Computer Science staff will support SISTA’s financial, academic and technical requirements.  Other participants in SISTA will come from MIS, ECE, Optical Sciences, Cognitive Science, Sociology, Systems and Industrial Engineering, Ecology and Evolutionary Behavior, the BIO5 institute, the Statistics GIDP, and Biosphere II.

So what are the advantages of assembling the Geekodrome? For one thing, it means advances in understanding information can be applied to other subjects:

Information Science predates computers, but without computers much of the theory we associate with Information Science would not have been developed, and it certainly would not have been applied to produce the economic impact of recent decades.  But because computers and large datasets became universally available quite recently – more recently than many academic departments formed – information scientists are developing essentially similar ideas in departments that rarely interact.  For instance, viral marketing strategies in business, percolation in physics, models of human epidemics, and the spread of computer viruses are essentially similar ideas, analyzed with essentially the same models, from distinct academic communities.

The idea of giving a wide range of researchers a basis in information science is also addressed later in the proposal:

Faculty wish [sic?] be members of both SISTA and their home departments, and for good reasons: For many researchers, information science and technology are means, not ends. A sociologist who develops fundamental theory in social networks is a sociologist first, an information scientist, second. These researchers often are the “new wave” in their respective fields and should continue to lead the information revolutions in their departments.   

As science stumbles into Chris Anderson’s Petabyte Age, an era when massive datasets and the supercomputers that love to crunch them are creating powerful new research methods, UA ought to be ready to apply them, and this school sounds like a good way to do just that. But don’t listen to me!—listen to George Dyson, writing in Edge earlier this year:

The sudden flood of large data sets and the opening of entirely new scientific territory promises a return to the excitement at the birth of (modern) Science in the 17th century, when, as Newton, Boyle, Hooke, Petty, and the rest of them saw it, it was “the Business of Natural Philosophy” to find things out.

UA definitely ought to be preparing students and scholars for that.

The paper goes on to point out four potential benefits of the school. First, a school of Information Science is a (relatively) novel idea that would stand out among the “old-style Informatics programs” that dominate the field. Second, it creates a chance “for researchers who work alone or in small enclaves all over campus to join the SISTA community, to teach courses and write proposals and conduct research with people they never see.” Wait—wasn’t this a virtual school? Maybe they can Skype between cubicles. Third, the school allows UA “to apply information science to outreach and the STEM pipeline,” a conduit that sounds like it pumps crude oil through the Caucasus, but actually pumps eager math and science students through our university. Fourth, it “can both save and earn money for the UA.”

Savings and curriculum are noteworthy strengths of this plan. The proposal suggests cutting costs not just by sharing coffeemakers, but by revising the curriculum to actually benefit students. This section of the paper is worth quoting in full:

Students don’t realize that stochastic processes underlie models of gene mutation, language translation, predictive compilers, musical and space grammars, machine learning and data mining. They don’t realize that conditional independence is an essential organizing principle for networks, whether they are social networks, causal models, or protein-protein interaction networks.  They don’t realize that recall, precision, ROC curves, sensitivity and positive predictive value are closely related concepts given different names by different fields.  They don’t realize that some heuristics used in computer vision were invented by perspective painters five centuries ago.   

Students don’t understand the ubiquity of ideas in information science because we don’t teach them to see the connections.  The SISTA curriculum is based in six entirely new core courses that will teach the great concepts, techniques and issues in information science (see Appendix B). These will be followed by thematic courses that provide more depth, though still in a thoroughly interdisciplinary way.  For example, sequences are important data representations in intensive care, earthquake prediction, molecular genetics, marketing, and many other fields.  A thematic course on sequences would teach the common representation and inference methods for sequences, including parsing, ngram models, alignment algorithms, matching and predictive methods, and so on.  After taking core courses, and a selection of thematic courses appropriate to their majors, most undergraduates will complete their  coursework with more specialized courses from their major departments.   

The SISTA curriculum is intended to reduce duplicative teaching and encourage cooperation between units.  For instance, the UA has, but probably doesn’t need, at least 23 introductory programming courses, 13 courses on data management, and more than 30 courses in probability and statistical methods.  SISTA will provide the opportunity for information scientists all over campus to review, coordinate and, ideally, retire some of their course offerings.

Scrapping these redundant course offerings is an excellent idea. If you’re a social-science major like me, you’ve probably taken a statistics course of some kind. In my case, it was ECON 339, a big, basic class covering the fundamentals of stats. However, when I took it, there was nothing particularly econ-oriented about it. Like most simple statistics courses, most of the examples involved flipped coins, rolled dice, or student test scores. Some of these course superfluities are the natural result of student demand for “practical courses”—pure theory inevitably generates the standard complaints of “I’ll never use this in the real world” and “I won’t understand it unless I see it applied.” Take that advice seriously, and it’s no wonder we have 30 different statistics courses, many of which no doubt use the same hackneyed illustrations of probability principles. I’d much rather be learning about “space grammars,” whatever those are. 

All in all, this is an excellent, diligent plan and one of the clearest papers so far.

Photo via wearscience.


One Response

Subscribe to comments with RSS.

  1. crismark4 said, on 23 October 2008 at 10:09 pm

    hehe.. This page about office cubicles is good too.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: