Meaning: A good educator opens his students' minds to new ideas and pushes them to do things they haven't done before.

This is part of the David E. Kelley School of Advising series.

--
Banazir

In the first of my recent posts on the technological singularity, I remarked on something I called the "Ramanujan Syndrome". Modern biographers of the famous Tamil mathematician Srinivasa Ramanujan have often noted that his spiritual devoutness was over-romanticized, as contemporaries such as G. H. Hardy observed. Equally, however, there has in recent decades been a tendency to view his lack of formal training before the age of about 25 (in 1914) as a distinction, almost glamorous.

( M. T. Iyengar, editor of the Journal of the Indian Mathematical Society, on Ramanujan )
( Ramanujan's work with Hardy, 1914 - 1919 )
( My remarks )

What do you think?

--
Banazir

How do you keep people from grabbing you when you are swamped? Do you sit down with them and have a quiet, reasoned discussion, explaining why you are sometimes too busy for them? Do you dash on, yelling "sorry, not now!" or "gerroff!"? If so, do you explain later by e-mail, in person, or not at all? Do you stop and stare, say it's rude, threaten mayhem?

Most important: Does it work? Do you ever get grabbed by the same person again? If so, is it only in exigent situations or also for non-emergencies that they grab you?

--
Banazir

حكمة [transliterated hikma] (Arabic, noun) - "wisdom"

From Wikipedia:

The House of Wisdom (Arabic: بيت الحكمة‎; Bait al-Hikma) was a key institution in the Translation Movement - a library and translation institute in Abbassid-era Baghdad, Iraq. It is considered to have been a major intellectual center of the Islamic Golden Age. The House of Wisdom acted as a society founded by Abbasid caliphs Harun al-Rashid and his son al-Ma'mun who reigned from 813-833 CE. Based in Baghdad from the 9th to 13th centuries, many of the most learned Muslim scholars were part of this excellent research and educational institute. In the reign of al-Ma'mun, observatories were set up, and The House was an unrivalled centre for the study of humanities and for sciences, including mathematics, astronomy, medicine, chemistry, zoology and geography. Drawing on Persian, Indian and Greek texts—including those of Pythagoras, Plato, Aristotle, Hippocrates, Euclid, Plotinus, Galen, Sushruta, Charaka, Aryabhata and Brahmagupta—the scholars accumulated a great collection of knowledge in the world, and built on it through their own discoveries. Baghdad was known as the world's richest city and centre for intellectual development of the time, and had a population of over a million, the largest in its time. The great scholars of the House of Wisdom included Al-Khawarizmi, the "father" of algebra, which takes its name from his book Kitab al-Jabr.

--
Banazir

As some of you know, we've been revamping our system of preliminary exams for the Ph.D. in Computer Science from more of a pure comprehensive exam system to one of written comps (breadth exams) plus oral quals (depth exams or "research proficiency exams"). The first offering of the breadth exam was in fall, 2006.

Is this really the best system, though? I've heard recently that many top universities have moved completely away from comps and quals, to a "progress review" system. We had heard of this practice, first used in CS graduate education by Carnegie Mellon University's School of Computer Science, where it was called "Black Friday". I set out to examine this alternative system and consider its pros and cons.

( Pros of the Black Friday system )
( Peter Lee on Black Friday at CMU )
( Jeanette Wing on Black Friday at CMU )
( Cons of the Black Friday system )

Conclusion: My view is that Black Friday can be a good thing if we have basically turned exams into formalities. It can also ease the strain to something more equitable if we've made things too competitive (which was the rationale for moving away from pure comps in the first place). I don't think standards are intrinsically "too challenging" or "not challenging enough"; rather, I feel that we need to let students "rise to the level of expectations", in the words of Jaime Escalante. We also need to feel some personal accountability for showing them how to achieve this objective. As long as they think they are just jumping through hoops to satisfy us "as a faculty", there will continue to be people who just fall through the cracks. That includes passing without really getting to the next level! In the end, the tasks and milestones we set should challenge students to get better, rather than strike them immobile with the fear of failure. It should be not "up or out", but excelsior.

--
Banazir

Is your institution accredited by the Accreditation Board for Engineering and Technology (ABET)? How about North Central? What do you think of these agencies, and of accreditation in general?

When I was at Johns Hopkins as an undergrad, the Mathematical Sciences department (now Applied Mathematics and Statistics) was just going through ABET accreditation for the first time. Since I came to K-State in 1999, we've had two ABET visits and two North Central visits. As you may know, some of the top American universities, such as MIT, Stanford, and CMU (which happen to occupy the top three slots among CS Ph.D. programs, according to the Computing Research Association, U.S. News and World Report, and several other ranking agencies) don't even bother with accreditation, on the principle that their reputations stand for themselves. While I think this is true, I was just curious as to people's stances on the matter.

--
Banazir

Meaning: Black Friday, people. It's not just for CMU anymore...

This is part of the David E. Kelley School of Advising series.

--
Banazir

Are members of your family devout believers in some religion or system of belief? Would you call them "fundamentalists", either in the literal sense of returning to the original scriptures and doctrines of a system of belief, or the colloquial sense of being very hidebound adherents to the old-time religion?

I would call my folks Confucian fundamentalists. In some sense, they believe in the Taoist ideals of the Doctrine of the Mean, eschewing extremism and fearmongering in the name of theocratic authority. They believe in the principle that mankind should live in harmony with society, seeking first to set oneself, then one's family, then one's village, and finally one's nation right. They hold the humility, the love of learning, and humanity highest among virtues.

( The Five Bonds: Civic, Filial, Marital, Familial, Social )
( Analects )

( A few comments and critiques )
( Seeming fair and feeling foul: Analects I.3 )

--
Banazir

Poll #1296348 Board Poll
Open to: All, detailed results viewable to: All, participants: 30

Do you like white boards?

Do you like blackboards?

Which feature of white boards do you dislike the most?

Which feature of black boards do you dislike the most?

If you had to switch to JUST ONE, which would it be?

--
Banazir

David E. Kelley is the creator and producer of many television shows in the "legal drama" genre, such as L. A. Law, The Practice, Ally McBeal, and Boston Legal. He also created a show I liked very much, Boston Public.

Now, those of you who know me well in person, or even online, probably know that I am given to hyperbolic expressions and analogies. Many of you have scratched your head at some allusion or other and said, "whatchu talkin' 'bout, Bana?". Well, now that we have this nifty little tool called YouTube, I thought I would set some of these questions to rest with an occasional videoblog entry in what I call the "David E. Kelley School of Advising" series. This is really what plays through my inner mind when I make allusions.

For example, what is the David E. Kelley School of Advising? Well, on Ally McBeal, the introspections and reactions are shown onscreen in various humorous ways, such as this:




That's about what I used to look like getting ready for a defense or presentation (and probably still do, sometimes).

So now, when I feel jubilantly bloodthirsty, I don't have to write ROMA VICTOR!, and when I'm about to go ballistic, I don't have to express my rage with ALDKJHAFKL;ASAFJAFALJ. I can just post one of these! It'll be my "Darmok".

--
Banazir

• Who needs it: Architects, landscape architects, urban planners, interior designers, and furniture designers are the core collaborators for this discipline. It is also relevant to design-oriented disciplines from textiles to mechanical engineering (CAD, CAE) to industrial and manufacturing systems engineering (CAM). I have talked with people in Civil Engineering who have shown me the relevance of this area to CE, Architectural Engineering and Construction Science, Materials Science, and even areas of the physical and life sciences where design technology is needed. The physics and materials science aspects are also related to nondestructive evaluation (NDE).


• Why we need it: Development of tools is a primary rationale, but there is also staying timely. We need to remain receptive and adaptive to new fields. For example, we have lots of people with Human-Computer Interaction (HCI) interests, and we have human factors researchers with expertise in certain kinds of ergonomics, but there just isn't a place for a user interface ergonomist (for example) to earn tenure. At Virginia Tech or UIUC, there are always a few people who have HCI as a core area, and there is always a center such as the UIUC Beckman Institute where they can cross disciplines to get cooperative efforts funded and published. Here, we have to do a little more than ask.


• When we need it (and how long we've needed it): We need it ASAP, and this has been the case since the late 1990s. We have some visualization infrastructure in the College of Architecture - an Elumans VisionDome and more at the Krider Center operated by Jeff Head - but it yet hasn't quite hit the big time in terms of full utilization. Similarly, Charlie Zheng in Mechanical and Nuclear Engineering (MNE) uses visualization in his research, but there isn't yet a college-level visualization center in Engineering or a university-level one. I've operated a visualization lab in our department for eight years, but aside from massforge, there hasn't been a long term project that it supports. By contrast, top research universities such as MIT have CSAIL (a recombination of the Laboratory for Computer Science and the Artificial Intelligence Lab), the Media Lab, etc. The computer graphics course is jointly offered by CS and Architecture.


• How bad we need it: A visualization lab is just the beginning. Planning informatics - from simulation research and tools development - would help us advance the state of the field in intelligent systems. So would a research and teaching program in informatics applied to design. We need these as much as we need work on optimization and scheduling. I believe that this is one of the most eminently fundable areas in all of informatics.


• Where it should live (i.e., to what college it should belong): Seeing as it would be primarily an offshoot of Landscape Architecture and its urban planning program (Regional and Community Planning), I think the College of Architecture, Planning, and Design would be its natural home.

--
Banazir

• Who needs it: Biologists, especially molecular biologists, but also ecologists and population genetics people; biochemists; pharmacologists. In Arts and Sciences, entomologists too. Agricultural researchers, especially plant pathologists and physiologists and animal scientists, but also including other agronomists, horticulturalists, etc.. Pathobiologists and clinical researchers in the College of Veterinary Medicine. People in food safety and at the National Agricultural Biosecurity Center (NABC). Computer scientists from intelligent systems to computer architecture to programming language theorists, as an application area; statisticians, and mathematicians for similar reasons. Chemoinformatics and computational physics, for tips of the trade.


• Why we need it: See the application domains above for specific reasons. As with Library Science, it's also a sign that you have arrived on the scene if you can tie together some of these areas. Also, as with BME, it's a good way to become more marketable, garner NIH and NSF funding, and just gain prestige as a "Research One" university.


• When we need it (and how long we've needed it): We've been ready for no more than five years, if that. I was part of the first wave of bioinformatics people who organized circa 2001 and really started making the big postdoc hires only in 2003 and the big faculty hires only in 2006. Doina Caragea is our first fully-fledged bioinformatician in CS, though people such as Sue Brown in Biology, Clare Nelson in Plant Path, Steve Welch in Agronomy, Sanjoy Das in EECE, and myself work on various bioinformatics projects.


• How bad we need it: Once again, it's a question of opportunity costs, but this time I can put a name to a face: NSF has had the Frontiers in Integrative Biological Research (FIBR), Arabidopsis 2010, and Biological Databases and Informatics (BDI) programs; we have our Ecological Genomics and Arthropod Genomics programs from NSF and Targeted Excellence; and then there are all the NIH R01 projects.


• Where it should live (i.e., to what college it should belong): I, personally, would like to see it in Agriculture, but perhaps the Division of Biology would be appropriate as they also have a strong program in molecular genetics. As with Library Science and Linguistics, it already exists in "spread out" form, and this is more of a need for principled, consolidated growth and program development.

--
Banazir

• Who needs it: Everyone! No, really: everyone in research. Librarians as practitioners of the science and purveyors of problems, data, and participants. Education folks in both research and pedagogical training for related reasons. CS and statistics researchers for the information retrieval (IR) and information extraction (IE) angle. Human-computer interaction (HCI), human factors, and ergonomics people in IMSE and Psych. Modern languages people and anthropologists to study the social impact, translation. Linguists for the natural language processing aspects (see the first department I advocated for). Applied mathematicians for the fundamental probability, graph theory, and combinatorics. Scientists as end users and hence providers of user data. People from the humanities for critical review of texts. Journalism and political science researchers for their expertise and interests in specialized libraries. Medical and legal researchers from other Kansas Regents universities.


• Why we need it: See above for domain-specific reasons. I also think that a Library Science department is a good sign that a large public university has arrived when it comes to the interfaces between CS and the university library, among different HCI researchers across colleges, and among different information science disciplines from cognitive science to the various flavors of intelligent systems. The more integrated we are, the better we can talk to one another, and the more we can leverage interdisciplinarity. We're not even close yet to having something like a Library and Information Science department, much less Virginia Tech's centers and UIUC's top-ranked Graduate School of Library and Information Sciences (GSLIS), but you have to start somewhere. As Mehdi Harandi said when he was on my Ph.D. committee: "The majority of lost opportunities for interdisciplinary collaboration can be blamed on people's unwillingness to walk a block". I find it's also a good sign when the left hand knows what the right is doing - for example, when different academic computing units (such as our IT and Academic Computing group, Computing and Telecommunications Services, and our High-Performance Computing infrastructure group) unite in common cause.


• When we need it (and how long we've needed it): Decades, at least. Twenty years is probably a conservative estimate. This is a sensitive issue, I guess, because it's largely a function of how well-equipped and willing our university library (a college-level unit) is to do research and to teach courses in the discipline. I don't pretend to know the answer to these questions.


• How bad we need it: Again, as a function of opportunity cost from untapped markets, we need it to the tune of a few million dollars a year, conservatively. Whether that's 1 million or 10 million is debatable, but the opportunities abound, and potential for synergy is seldom higher across our campus. The right energy is there, too; the real issue is how efficient we can be about applying it and having all the pieces fit together.


• Where it should live (i.e., to what college it should belong): I, personally, would like to see it in the Library, but maybe Arts and Sciences would want it more. It can fit in several programs; we have only to create it.

--
Banazir

• Who needs it: We don't have a medical school, but the University of Kansas (KU) does, and we have the stronger computer science and overall engineering research programs here. Electrical, Computer, Mechanical, Industrial, Civil, and Chemical Engineering could all benefit from contact with BME. That's just our college; then there's Anatomy and Physiology in Vet Med, Physics, Biology, Statistics (for the OR, optimization, and control theory), Math, Biochem, Chemistry, in roughly that order. Business Administration, MIS, even Architecture and Design for the visualization aspects. Maybe the Genetics program in Plant Path. Plus, you don't have to have a medical school to have a medical library.


• Why we need it: To have research in telemedicine, applied sensor networks, and embedded systems - all very fundable areas; to expand into applied biophysics, microbiology, diagnostic medicine; to build on our core competencies in robotics, control systems, and optimization.


• When we need it (and how long we've needed it): Many engineering faculty, staff, and students would doubtless disagree with my assertion that we've been ready for a BME department for the last 10-20 years at least, but I'm speaking from what I saw at Hopkins, which has a top BME department and pretty good EE and CS departments. Also, UIUC didn't have BME when I was a grad student there, and now it does. And if that doesn't convince you, well... look! Vitruvian Man icon! Sparkly!


• How bad we need it: As a function of opportunity cost from unmet needs and unclaimed available funding, I would say we need it more than anything else. In terms of sheer intellectual benefit I put Linguistics first, but that's more an osmosis and head count thing.


• Where it should live (i.e., to what college it should belong): Engineering would be best equipped for BME, though I could see Vet Med housing a biomedicine program with certain telemedical and prosthetic components. I don't know enough about CVM to say, but from my experience: they are surprisingly well-funded by NIH, but in terms of untapped potential and staying power in bioinformatics, it's going to take major input from Engineering, Arts and Sciences, and Agriculture. Plant sciences gets short shrift in the BME arena, but if one of you has the perfect Bionic Ent idea, now's your chance to post a comment about it without looking like a kook.

--
Banazir

• Who needs it: People in speech communications, linguistic anthropology, modern languages, psycholinguistics, computational linguistics, and many other domains involving human language would benefit from collaborations with full-time (first career) linguists.


• Why we need it: We would be able to make advances in translation science, including machine translation; information extraction; first, second, and foreign language acquisition; cognitive modeling in speech production and linguistic memory; analysis and curation of endangered languages; and many other such fields.


• When we need it (and how long we've needed it): We've needed a linguist for over a decade, when Speech Communications lost the last "full time linguist" to retirement. Since then, Dr. Harriet Ottenmeyer has retired from Anthropology, a department that has hired a linguistic anthropologist (Dr. Tiffany Kershner), and we've talked with linguistic anthropology faculty at other universities (Dr. Arienne Dwyer at KU, for example), but there isn't a linguistics degree program, much less a department.


• How bad we need it: The Targeted Excellence proposal brought together at least eight people who could start working with one or more linguists in various areas.


• Where it should live (i.e., to what college it should belong): Most likely, a Department of Linguistics ought to be part of the College of Arts and Sciences, possibly sharing some faculty with Modern Languages via joint appointment. I don't know the structure of the ML program well enough to know whether spinoff adjunct appointments are feasible or desired by the faculty and administrators.

--
Banazir

Working on a Provost's Targeted Excellence proposal on Human Language Technologies (HLT) and gathering together the people with the requisite expertise to make it all work really brought home a point for me: we are short quite a few departments. As a Socratic exercise, I thought I would tell you which five I think are the most markedly and urgently absent.

( The list - move on to the next post if you want each revealed in its own time )

--
Banazir

OK, if definitely seems as if the CIS 490/730 (Principles of AI) midterm exam was a bit long. People generally skipped half a problem to one full problem out of 5.

Edit, 06:00 CDT Mon 24 Oct 2006 - Here is the exam, and here is the solution.

--
Banazir

Apropos of our continuing discussion treating adult students as such: how much do students really know what is good for them, as far as workload and expectations are concerned?

I ask because I'm sure that I've earned the reputation (for better or for worse) of having high expectations. On the plus side, this means I have acceptably high standards and am less likely to turn out "Jaywalker bait" into the CS/IT industry. I call this "keeping our name out of the dumb columns". On the minus side, I've been called a slave driver (though only in Chinese to my face, when the speaker didn't think I could understand; that was amusing).

rsmit212 showed me a comic strip once that concluded, "where knowledge is the commodity, the customer is always wrong". That's a clever notion, but how true is it?

I'm really looking for discussion, BTW, not just validation or critique. Some of the best insights I have gotten from my blog have been through dialogue, sometimes debates between second and third parties that I am only involved in as a bystander or facilitator.

--
Banazir

My first evening hour exam was Fri 06 Oct 2006 - I moved it from class time (2:30 - 3:20pm) to afternoon (5:30 - 6:45pm) under the assumption that some or most could use an extra 25 minutes.

Afterwards, students told me I was right. "It would have been possible but a tight squeeze to finish it in 60 minutes and extremely hard in 50, but with 75 we had about 10 minutes left to clean up and check our work."

I've heard that exams should be written so that students get 3-5 times as long as it took the instructor to solve the problem. I usually allow about twice as long (60 minutes for an exam I work in 30 minutes, 75 for one that takes me 35-40 minutes), and it's been a tight squeeze.

What's your experience been, as a student or as an instructor? (Please say what you took or taught.)

--
Banazir