?

Log in

No account? Create an account

Previous Entry | Next Entry

Grad Council: Persistence of Knowledge

Just came back from a KSU Grad Council meeting this afternoon.
Some of the discussions reminded me of an endemic problem that I have, which I like to call persistence of knowledge. For example, over the last 4.5 years, I've spent time bringing at least 3 generations of B.S. and M.S. students up to the task in areas of:

  • Machine learning

  • Pattern recognition

  • Graph theory


The turnaround time for grad students, however, ranges between 1 and 4 years for my students, with the average around 18 months for M.S.E. and M.S. students. And when they graduate, they take their earned expertise with them, leaving me with a fresh generation to train. This is part and parcel of the academic circle of life - as Dr. Michael Loui writes in his talk How to Choose A Thesis Advisor: "although you start as an apprentice, ideally, you should end as a colleague".1 It reminds me slightly, however, of the ending of Fahrenheit 451 by Rad Bradbury:

"What have you to offer?"
  "Nothing. I thought I had part of the Book of Ecclesiastes and maybe a little of Revelation, but I haven't even that now."
"The Book of Ecclesiastes would be fine. Where was it?"
  "Here," Montag touched his head.
"Ah," Granger smiled and nodded.
  "What's wrong? Isn't that all right?" said Montag.
"Better than all right; perfect!" Granger turned to the Reverend. "Do we have a Book of Ecclesiastes?"
  "One. A man named Harris in Youngstown."
"Montag... Walk carefully. Guard your health. If anything should happen to Harris, you are the book of Ecclesiastes. See how important you've become in the last minute!"


Anyone have any ideas for maintaining sustainable expertise? hfx_ben, who's run big software development projects before, and f00dave, who's finishing up a major long-term paper, can probably speak to this particularly well.

Other notes from the Grad Council:

  • We are having a Graduate Forum again soon, with the abstract submission deadline set for 15 Mar 2004.

  • We are about to institute an application fee for domestic students for the first time, and raise the application fee for international students, in order to defray graduate recruiting and admissions costs (namely, to maintain and expand clerical support for the Graduate College, develop software for applications processing, and also to put together some basic recruiting materials).

  • The Provost's Targeted Excellence might be sustaining a funding hit. All indications are that the expected number of 3 or more funded proposals (out of 14 that made the preproposal cut) will come to pass, but at a reduced level.

  • Compliance with the Student and Exchange Visitor Information System (SEVIS) was achieved last year, and acclimation to SEVIS system is almost a reality.



1 ObFoE: There's some relevance here for the Jedi (and perhaps the Sith), as well. Opinions?

--
Banazîr

Comments

( 9 comments — Leave a comment )
hempknight
Mar. 8th, 2004 03:16 pm (UTC)
Um...
Do you want to keep the people or just the knowledge/information? In the first case just throw money at them.

For the other, a knowledge management system of sorts could be the answer. Decentralized in nature, like an intranet/extranet architecture where you basically enter whatever you deem important from wherever you want. In other words, if someone discovers something in the middle of the night they can enter it in the knowledge base from their bed if need be. Decentralize it and even people who have gone away could still have access to it and feed it as it were.

I've created quite a few of those already but never for educational purposes so I'd assume that data model would differ from say, a construction company, but the idea remains the same basically. My thesis was about a similar thing actually.

--
Danger is my middle name
banazir
Mar. 8th, 2004 08:11 pm (UTC)
People or knowledge?
Do you want to keep the people or just the knowledge/information? In the first case just throw money at them.
I want to keep the people, but I can't keep them indefinitely in an academic environment.
Hiring istari_ala is a special case.
Generally, if it's an M.S. student, I can keep them around at most one year after graduation as an associate scientist (and that would require tons of money, as you said). If it's a Ph.D. student, I could offer a postdoc, but they probably would, and shouldn't, accept.

For the other, a knowledge management system of sorts could be the answer. Decentralized in nature, like an intranet/extranet architecture where you basically enter whatever you deem important from wherever you want.
Like this Wiki for geckies? Or this one for RoboSim?
Yeah, you just described istari_ala's first job. Thanks. :-)

I've created quite a few of those already but never for educational purposes so I'd assume that data model would differ from say, a construction company, but the idea remains the same basically. My thesis was about a similar thing actually.
Tell us more, please!

Thanks for your comments.

--
Banazir
hempknight
Mar. 9th, 2004 01:45 am (UTC)
Re: People or knowledge?
I want to keep the people, but I can't keep them indefinitely in an academic environment.

Throwing money at them always does the trick, trust me.

Like this Wiki for geckies? Or this one for RoboSim?
Well, a bit more focused than that. Set stricter rules (from within the system itself) on what goes in and what goes out. You'd first need to research and chart all your information flows and your communication needs before you can set those rules and then decide on formatting and then...blahblah...boring...boring...blahblah...boring.

The main problem in KMSeseses is always how to reduce data pollution. Because data turns into information then and only then when the user of the said system finds it useful.

Am I boring you yet? Feel free to stop me anytime.

Yeah, you just described istari_ala's first job. Thanks. :-)

You're gonna get her to create a KMS all by her lonesome because of ME? She'll hate me forever and ever because of it!!! Must you ruin everything???

Tell us more, please!

About my thesis? I'd rather not. Still occasionally gives me nightmares. Working full time and graduating at the same time is something I care not to repeat.

I was presented with a B2B problem of project management. I was to solve all the common problems which go with decentralization of projects (=outsourcing). Basically the boss here wants to know what his butt-monkeys are doing half way across the globe. So I set out on a voyage of self discovery and decided that such thing is possible but not at the company which presented the case. I came up with a XML-based EDI system which controls the project flow pretty much autonomously. Users get the output they want even when they're not asking for it. Through e-mail, fax AND mobile phone. It is a Central Neural-Net Processor of sorts, a learning computer. If they actually built it, it would have taken over the world most certainly. Sometimes things work out for the better.

--
Danger is my middle name
zengeneral
Mar. 8th, 2004 04:20 pm (UTC)
Enjoy insanity
    I am assuming this is in reference to the mind breaking down since books preserve knowledge quite nicely. If this is the case then there are two solutions:

Solution 1:
        I imagine, in my world, the solution is to put reason, rationality, logic and all other fun things in a bag. Then throw this bag outside and stomp on it. This way, when you are nice and insane, you can do whatever you want; what you do would make you an expert at doing it. Course, this might not be nearly as fun as what you want to do with all those fun things that are broken in the bag.

Solution 2:
        Nevermind, I forgot it….
banazir
Mar. 9th, 2004 12:05 am (UTC)
Re: Enjoy insanity
I am assuming this is in reference to the mind breaking down since books preserve knowledge quite nicely.
No, no, this was in reference to my having to reteach things that my senior students knew, and which were once part of the "collective lore" of my research group that only elicits blank looks after the knowledge has "passed beyond the ken" of the current crop of students.

If this is the case then there are two solutions:
Solution 1:
[trasking logic and rationality]

Uhhhhh...
Oooooookaaaay...

Solution 2: Nevermind, I forgot it….
It's an open-book exam.

--
Banazir
(no, really, it is!)
gondhir
Mar. 9th, 2004 12:14 pm (UTC)
Download your students' brains to your servers prior to them leaving.

Just be sure to get a bigger/more stable RAID array for Fingolfin first though...
banazir
Mar. 9th, 2004 12:21 pm (UTC)
Must... resist...
Download your students' brains to your servers prior to them leaving.
Just be sure to get a bigger/more stable RAID array for Fingolfin first though...

:choked grin:

Must not say it...
Must not say it...
Must not say it...

Oh, traskit!
It's no problem, they'll fit with gigs to spare!!!

Ahhhh...

--
Banazir
gondhir
Mar. 12th, 2004 07:35 pm (UTC)
Re: Must... resist...
Just as long as you don't try it with your (future) employees? ;)
banazir
Mar. 12th, 2004 07:40 pm (UTC)
Re: Must... resist...
Just as long as you don't try it with your (future) employees? ;)
They have no more built the RAID that can contain istari_ala's brain than they have built the cage that can contain teunc.

Or FoE, for that matter.

--
Banazir
(your brain, OTOH... ;-))
( 9 comments — Leave a comment )

Latest Month

December 2008
S M T W T F S
 123456
78910111213
14151617181920
21222324252627
28293031   

KSU Genetic and Evolutionary Computation (GEC) Lab

Teunciness

Breakfast

Science, Technology, Engineering, Math (STEM) Communities

Fresh Pages

Tags

Powered by LiveJournal.com
Designed by Naoto Kishi