Neural Network Customer

new topic     » topic index » view thread      » older message » newer message

For those interested in the Neural Network, this is what my brother sent me
concerning their needs and interests:

"The NN would have to be able to apply higher dimension linear regression to
large sets -- at high speeds."

"Basically it will be given about 200,000 8 dimensional arrays and it will
have to draw a 7dimensional line separating clusters into distinct types
(they'll have High overlap) ... the separation will then go to a second
data input (same stats) and will have to sort cells into one type or
another. "

--Noah's comment:
If i remember correctly, what he's talking about here is seperating cells into 2
groups.  There are eight "factors" which can describe a given cell, some which
determine if the cell is abnormal.  The problem is, there are 3 types of cells,
"normal" run-of-the-mill cells, "stem" or really good cells, and "cancer" or
really bad cells.  They want stem cells, and don't want cancer cells, but the
same data which defines a stem cell appears to also define a cancer cell.  They
have to find a "line" (the "line" is a seven-dimensional structure) which
seperates the eight dimensional data set into 2 groups.
--

"The entire process must occur within 10 to 15 minutes, the sorting will
take place at around 40000 cells/sec or much more. "

--Noah's comment:
The code will probably be hard-wired in some fashion to a device which actually
sorts the cells based on the information it receives.
--

"We would be willing to pay but we're broke. Compiled code or source code is
fine (we a have a number of old, old, old timers here [like old
machine/algol 58 guys]). "

--Noah's comment:
There's a bunch of _really_ old programmer's here.  For instance, I beleive Dr.
Hokanson's been programming for ~50 years.
--

"The system it would run on is probably going to be hard-wired with the
program, so go with win98, and anything sgi, sun, macos and cray in that
order. "

--Noah's comment:
i.e.  they can use almost anything.  I think they have a PGP11? running some
equipment there, and a few other devices all connected in parallel.  if it's
coded, they can probably hook it up.

"What would be really nice would be sgi, since we're getting to use those
machines this summer. "

"The above task is the only one. "

"Here are some references on the project: "

"Leary, J.F., Hokanson, J. A., McLaughlin, S.R. "High Speed Cell
Classification Systems fo REal-Time Data Classification and Cell Sorting"
SPIE (Journal of the Optical Society of America) 2982: 342-352, 1997. "

"Leary, J.F., McLaughlin, S.R., Hokanson, J.A., Rosenblatt, J.I. "High Speed
Real-time Data Classification and Cell Sorting Using Discriminant Functions
and Probabilities of Misclassification for Stem Cell Enrichment and Tumor
Purgin." SPIE (Journal of the Optical Society of America) 3260: 274-281,
1998a. "

"there is also another paper, but I don't know if it's published: "

"Leary, J.F., McLaughlin, S.R., Reece, L.N., Rosenblatt, J.I., Hokanson,
J.A. Dept. of Internal Medicine, Division of Infectious Diseases, Molecular
Cytometry Unit; Dept. Preventive Medicine & Community Health University of
Texas Medical Branch, Galveston, Tx 77555. "Real-time Multivariate
Statistical Classification of Cells for Flow Cytometry and Cell Sorting: A
Data Mining Application for Stem Cell Isolation and Tumor Purging." "

"they can get Dr. Hokanson by emailing him here:

jhokanso at utmb.edu "

"Any message must be RATHER succinct (he's extraordinarily busy and gets an
email, and I shit you not, every 2 minutes 24 hours a day [seriously, I
worked at his computer for 3 weeks and in one 50 minute period he recieved
78 emails])."

"Give the subject as dealing with "discussion of data-mining and neural
networks" "

that's all, folks,

Noah



<HTML><BODY
BGCOLOR=#FFFFFF>For&nbsp;those&nbsp;interested&nbsp;in&nbsp;the&nbsp;Neural&nbsp;Network,&nbsp;this&nbsp;is&nbsp;what&nbsp;my&nbsp;brother&nbsp;sent&nbsp;me&nbsp;concerning&nbsp;their&nbsp;<BR>needs&nbsp;and&nbsp;interests:
<BR>
<BR>"
<FONT COLOR="#000000" SIZE=3 FACE="Times New
Roman">The&nbsp;NN&nbsp;would&nbsp;have&nbsp;to&nbsp;be&nbsp;able&nbsp;to&nbsp;apply&nbsp;higher&nbsp;dimension&nbsp;linear&nbsp;regression&nbsp;to&nbsp;
<BR>
<BR>another.&nbsp;"
<BR>
<BR>--Noah's&nbsp;comment:
<BR>--
<BR>
<BR>
<BR>--Noah's&nbsp;comment:
<BR>--
<BR>
<BR>machine/algol&nbsp;58&nbsp;guys]).&nbsp;"
<BR>
<BR>--Noah's&nbsp;comment:
<BR>--
<BR>
<BR>order.&nbsp;"
<BR>
<BR>--Noah's&nbsp;comment:
<BR>
<BR>machines&nbsp;this&nbsp;summer.&nbsp;"
<BR>
<BR>
<BR>
<BR>
<BR>1998a.&nbsp;"
<BR>
<BR>
<BR>
<BR>
<A HREF=mailto:jhokanso at utmb.edu >jhokanso at utmb.edu </A>"
<BR>
<BR>78&nbsp;emails])."&nbsp;
<BR>
<BR>networks"&nbsp;"
<BR>
<BR>that's&nbsp;all,&nbsp;folks,
<BR>

new topic     » topic index » view thread      » older message » newer message

Search



Quick Links

User menu

Not signed in.

Misc Menu