Copy Link
Add to Bookmark
Report
Machine Learning List Vol. 5 No. 09
Machine Learning List: Vol. 5 No. 9
Sunday, April 25, 1993
Contents:
Machine Learning List: Vol. 5 No. 9
ML93 Applications Workshop
Flexible Discriminant Analysis by Optimal Scaling
AAAI Robot Building
Informatics Training at OHSU
1st Intl Conf on Intell Sys for Molecular Biology
The Machine Learning List is moderated. Contributions should be relevant to
the scientific study of machine learning. Mail contributions to ml@ics.uci.edu.
Mail requests to be added or deleted to ml-request@ics.uci.edu. Back issues
may be FTP'd from ics.uci.edu in pub/ml-list/V<X>/<N> or N.Z where X and N are
the volume and number of the issue; ID: anonymous PASSWORD: <your mail address>
----------------------------------------------------------------------
Date: Thu, 22 Apr 93 14:45:42 EDT
From: Pat Langley <langley@learning.siemens.COM>
Subject: ML93 Applications Workshop
CALL FOR PARTICIPATION
WORKSHOP ON FIELDED
APPLICATIONS OF MACHINE LEARNING
Amherst, Massachusetts
June 30-July 1, 1993
One of the central insights of AI is that expert performance requires
domain-specific knowledge, and work on knowledge engineering has led to
many AI systems that are now regularly used in industry and elsewhere.
The ultimate test of machine learning, the subfield of AI that studies
the automated acquisition of knowledge, is the application of its
techniques to produce similar results.
We believe the time has arrived for machine learning to concentrate
a substantial portion of its energies toward developing real-world
applications. The goal of this workshop is to familiarize participants
with existing applications of machine learning and to explore the
potential for additional ones in business and the public sector.
The workshop will revolve around invited talks by scientists who have
developed fielded applications of machine learning that are now in regular
use. Presentations will focus on the domain of application, the formulation
of the learning problem, any representational engineering required for
successful results, the community of users for the application, and measures
of the application's usefulness. We are encouraging speakers to downplay
differences among the specific learning methods employed, and to emphasize
the machinations that were needed to let their chosen technique solve their
particular problem. We also plan considerable time for open discussion.
To participate, simply check the space for Workshop B on the ML93
registration form. For more information about the meeting, contact one
of the workshop co-chairs, Pat Langley (langley@learning.siemens.com)
and Yves Kodratoff (yk@lri.fr), by electronic mail. This meeting is
co-sponsored by the U.S. Office of Naval Research, the Institute for
the Study of Learning and Expertise, and the IFIP Working Group on
Machine Learning.
------------------------------
From: trevor@research.att.COM
Date: Tue, 20 Apr 93 17:37 EDT
Subject: Flexible Discriminant Analysis by Optimal Scaling
Flexible Discriminant Analysis by Optimal Scaling
by Trevor Hastie, Rob Tibshirani and Andreas Buja.
This report covers a new class of nonparametric classification
procedures, obtained by generalizing Fisher's linear discriminant
analysis in a nonlinear way. The technique can be viewed as a
postprocessor to any form of multivariate regression of the
class indicator variables, and we argue that it is superior to
the traditional "softmax" in a number of ways.
A postscript version of the paper is available from NETLIB
Send the message to netlib@research.att.com:
send 93-13 from stat/doc
or else ftp to research.att.com, login as netlib, cd to stat/doc
and get 93-13
Trevor Hastie trevor@research.att.com
1-908-582-5647 (FAX) 1-908-582-3340
rm 2C-261 AT&T Bell Labs, Murray Hill, NJ 07974
------------------------------
Date: Tue, 13 Apr 93 08:58:15 PDT
From: NCAI <ncai@aaai.ORG>
Subject: AAAI Robot Building
Important addendum to AAAI-93 Conference Brochure regarding
the Robot Building Event.
Note that preregistration is mandatory for the Robot Building
Event, and there is a $50 lab fee. In addition, student scholarships
are available.
------------------------------
Date: Wed, 21 Apr 93 14:08:20 PDT
From: Bill Hersh <hersh@ohsu.EDU>
Subject: Informatics Training at OHSU
The Biomedical Information Communication Center (BICC)
at Oregon Health Sciences University (OHSU) has an opening
in its post-doctoral training program in medical informatics
for July or September 1993. With eight appointed and seven
adjunct faculty, along with a new 74,263 sq. ft.
state-of-the-art building, the BICC is one of the country's
leading institutions in medical informatics research. OHSU
has a major commitment to rural health education and outreach,
and the BICC mission to connect electronically 5,000 Oregon
health professionals on the nation's first statewide network
by the year 2000 will provide unique opportunities for
informatics fellows.
The program will train physicians, computer scientists and
others who are committed to a career in medical informatics.
The program has as its focus "end-user informatics," with
areas of concentration that include:
- Design and delivery of information resources and knowledge bases
- Organization and representation of health information
- Information retrieval
- Design and implementation of workstations for health professionals
- Health outcomes research
- Image analysis
- Informatics training and education
The primary focus of the program will be to provide a
structured research experience in one or two of the above areas,
along with course work in informatics at OHSU and in related areas
several nearby universities. Trainees will survey the field
broadly during their two to three year fellowship. They will
be expected to complete research projects, and upon completion
of their training, be able to describe their results clearly
in both oral and written form. The overall goals are to prepare
trainees to (a) direct their own medical informatics
research efforts at medical centers that actively embrace the
Integrated Advanced Information Management Systems (IAIMS) agenda,
or (b) take leadership positions in the growing number of hospital
and/or commercial efforts in medical informatics.
Qualifications for applicants include an M.D. (residency training
preferred) or a Ph.D. in biological science or an area relevant
to informatics. Financial support is available for U.S. citizens
or permanent residents only.
For more information, please contact:
Kent A. Spackman, M.D., Ph.D.
Associate Director for Academic Programs
Biomedical Information Communication Center
Oregon Health Sciences University
3181 SW Sam Jackson Park Rd.
Portland, OR 97201-3098
503-494-4502
spackman@ohsu.edu
------------------------------
Date: Thu, 15 Apr 93 11:51:23 -0500
From: Jude Shavlik <shavlik@cs.wisc.EDU>
To: ml@ics.uci.edu
Subject: 1st Intl Conf on Intell Sys for Molecular Biology
[For those attending the AAAI conf this summer, note that
this conference, which contains many papers involving ML,
is immediately preceding AAAI.]
PRELIMINARY PROGRAM AND REGISTRATION MATERIALS
First International Conference on
Intelligent Systems for Molecular Biology
Washington, D.C.
July 6-9, 1993
Sponsored by:
The National Institutes of Health,
National Library of Medicine
The Department of Energy,
Office of Health and Environmental Research
The Biomatrix Society
The American Association for Artificial Intelligence (AAAI)
Poster Session and Tutorials:
Bethesda Ramada Hotel
Technical Sessions:
Lister Hill Center Auditorium, National Library of Medicine
For more information contact ISMB@nlm.nih.gov or FAX (608)262-9777
PURPOSE
This, the First International Conference on Intelligent Systems
for Molecular Biology, is the inaugural meeting in a series
intended to bring together scientists who are applying the
technologies of artificial intelligence, robotics, machine
learning, massively parallel computing, advanced data modelling,
and related methods to problems in molecular biology. The scope
extends to any computational or robotic system supporting a
biological task that is cognitively challenging, involves a
synthesis of information from multiple sources at multiple levels,
or in some other way exhibits the abstraction and emergent
properties of an "intelligent system."
FACILITIES
The conference will be held at
Lister Hill Center
National Library of Medicine
8600 Rockville Pike
NIH, Building 38A
Bethesda MD 20894
Seating in the conference center is strictly limited, so
registrations will be accepted on a first-come, first-serve basis.
Accomodations, as well as a reception and poster session, will be
at the
Bethesda Ramada Hotel
8400 Wisconsin Avenue
Bethesda MD 20814
A special room rate has been negotiated with the hotel, of $92/day
(expires 6/21). Attendees must make their own reservations, by
writing the hotel or calling (800)331-5252 and mentioning the
ISMB conference. To participate in a roommate matching service,
e-mail opitz@cs.wisc.edu.
TRANSPORTATION
The two facilities are within easy walking distance, convenient to
the subway (Metro Red Line, Medical Center stop), and from there
to the Amtrak station. Nearby airports include Dulles, National,
and Baltimore-Washington International.
PROCEEDINGS
Full-length papers from both talks and posters will published in
archival proceedings. The citation is:
Proceedings of the First International
Conference on Intelligent Systems for
Molecular Biology (eds. L. Hunter,
D. Searls, and J. Shavlik) AAAI/MIT
Press, Menlo Park CA, 1993.
Copies will be distributed at the conference to registered
attendees, and will be available for purchase from the publisher
afterwards.
TALKS
Wednesday, July 7, 1993
================================================================
8:00-9:00am Continental Breakfast
9:00-9:15am Opening Remarks
9:15-10:30am Invited Talk
"Statistics, Protein Cores, and Predicted Structures"
Prof. Temple Smith (Boston University)
10:30-11:00am Break
11:00am "Constructive Induction and Protein Structure Prediction"
T.R. Ioerger, L. Rendell, & S. Surbramaniam
11:30am "Protein Secondary-Structure Modeling with Probabilistic
Networks" A.L. Delcher, S. Kasif, H.R. Goldberg, & W. Hsu
12:00-1:30pm Lunch
1:30pm "Protein Secondary Structure using Two-Level Case-Based
Reasoning" B. Leng, B.G. Buchanan, & H.B. Nicholas
2:00pm "Automatic Derivation of Substructures Yields Novel
Structural Building Blocks in Globular Proteins"
X. Zhang, J.S. Fetrow, W.A. Rennie, D.L. Waltz, & G. Berg
2:30pm "Using Dirichlet Mixture Priors to Derive Hidden Markov
Models for Protein Families" M. Brown, R. Hughey, A. Krogh,
I.S. Mian, K. Sjolander, & D. Haussler
3:00-3:30pm Break
3:30pm "Protein Classification using Neural Networks"
E.A. Ferran, B. Pflugfelder, & P. Ferrara
4:00pm "Neural Networks for Molecular Sequence Classification"
C. Wu, M. Berry, Y-S. Fung, & J. McLarty
4:30pm "Computationally Efficient Cluster Representation in
Molecular Sequence Megaclassification" D.J. States, N. Harris,
& L. Hunter
7:00-7:30pm Poster Setup
7:30-10:00pm Reception & Poster Session
Thursday, July 8, 1993
================================================================
8:00-9:00am Continental Breakfast
9:00-10:15am Invited Talk
"Large-Scale DNA Sequencing: A Tale of Mice and Men"
Prof. Leroy Hood (University of Washington)
10:15-10:45am Break
10:45am "Pattern Recognition for Automated DNA Sequencing:
I. On-Line Signal Conditioning and Feature Extraction for
Basecalling" J.B. Bolden III, D. Torgersen, & C. Tibbetts
11:15am "Genetic Algorithms for Sequence Assembly"
R. Parsons, S. Forrest, & C. Burks
11:45am "A Partial Digest Approach to Restriction Site Mapping"
S.S. Skiena & G. Sundaram
12:15-2:00pm Lunch
2:00pm "Integrating Order and Distance Relationships from
Heterogeneous Maps" M. Graves
2:30pm "Discovering Sequence Similarity by the Algorithmic
Significance Method" A. Milosavljevic
3:00pm "Identification of Human Gene Functional Regions Based on
Oligonucleotide Composition" V.V. Solovyev & C.B. Lawrence
3:30pm "Knowledge Discovery in GENBANK"
J.S. Aaronson, J. Haas, & G.C. Overton
4:00-4:30pm Break
4:30pm "An Expert System to Generate Machine Learning
Experiments: Learning with DNA Crystallography Data"
D. Cohen, C. Kulikowski, & H. Berman
5:00pm "Probabilistic Structure Calculations: A Three-
Dimensional tRNA Structure from Sequence Correlation Data"
R.B. Altman
5:30pm "Detection of Correlations in tRNA Sequences with
Structural Implications" T.M. Klingler & D. Brutlag
Friday, July 9, 1993
================================================================
8:00-9:00am Continental Breakfast
9:00-10:15am Invited Talk
"Artificial Intelligence and a Grand Unified Theory of
Biochemistry" Prof. Harold Morowitz (George Mason University)
10:15-10:45am Break
10:45am "Testing HIV Molecular Biology in in silico Physiologies"
H.B. Sieburg & C. Baray
11:15am "Identification of Localized and Distributed Bottlenecks
in Metabolic Pathways" M.L. Mavrovouniotis
11:45am "Fine-Grain Databases for Pattern Discovery in Gene
Regulation" S.M. Veretnik & B.R. Schatz
12:15-2:00pm Lunch
2:00pm "Representation for Discovery of Protein Motifs"
D. Conklin, S. Fortier, & J. Glasgow
2:30pm "Finding Relevant Biomolecular Features"
L. Hunter & T. Klein
3:00pm "Database Techniques for Biological Materials and
Methods" K. Baclawski, R. Futrelle, N. Fridman,
& M.J. Pescitelli
3:30pm "A Multi-Level Description Scheme of Protein
Conformation" K. Onizuka, K. Asai, M. Ishikawa, & S.T.C. Wong
4:00-4:30pm Break
4:30pm "Protein Topology Prediction through Parallel Constraint
Logic Programming" D.A. Clark, C.J. Rawlings, J. Shirazi,
A. Veron, & M. Reeve
5:30pm "A Constraint Reasoning System for Automating Sequence-
Specific Resonance Assignments in Multidimensional Protein
NMR Spectra" D. Zimmerman, C. Kulikowski, & G.T. Montelione
5:30-5:45pm Closing Remarks
POSTER SESSION
The following posters will be on display at the Bethesda Ramada
Hotel from 7:30-10:00pm, Wednesday, July 7.
[1] "The Induction of Rules for Predicting Chemical
Carcinogenesis in Rodents" D. Bahler & D. Bristol
[2] "SENEX: A CLOS/CLIM Application for Molecular Pathology"
S.S. Ball & V.H. Mah
[3] "FLASH: A Fast Look-Up Algorithm for String Homology"
A. Califano & I. Rigoutsos
[4] "Toward Multi-Strategy Parallel Learning in Sequence
Analysis" P.K. Chan & S.J. Stolfo
[5] "Protein Structure Prediction: Selecting Salient Features
from Large Candidate Pools" K.J. Cherkauer & J.W. Shavlik
[6] "Comparison of Two Approaches to the Prediction of Protein
Folding Patterns" I. Dubchak, S.R. Holbrook, & S.-H. Kim
[7] "A Modular Learning Environment for Protein Modeling"
J. Gracy, L. Chiche & J. Sallantin
[8] "Inference of Order in Genetic Systems"
J.N. Guidi & T.H. Roderick
[9] "PALM - A Pattern Language for Molecular Biology"
C. Helgesen & P.R. Sibbald
[10] "Grammatical Formalization of Metabolic Processes"
R. Hofestedt
[11] "Representations of Metabolic Knowledge"
P.D. Karp & M. Riley
[12] "Protein Sequencing Experiment Planning Using Analogy"
B. Kettler & L. Darden
[13] "Design of an Object-Oriented Database for Reverse Genetics"
K.J. Kochut, J. Arnold, J.A. Miller, & W.D. Potter
[14] "A Small Automaton for Word Recognition in DNA Sequences"
C. Lefevre & J.-E Ikeda
[15] "MultiMap: An Expert System for Automated Genetic Linkage
Mapping" T.C. Matise, M. Perlin & A. Chakravarti
[16] "Constructing a Distributed Object-Oriented System with
Logical Constraints for Fluorescence-Activated Cell Sorting"
T. Matsushima
[17] "Prediction of Primate Splice Junction Gene Sequences with
a Cooperative Knowledge Acquisition System"
E.M. Nguifo & J. Sallantin
[18] "Object-Oriented Knowledge Bases for the Analysis of
Prokaryotic and Eukaryotic Genomes"
G. Perriere, F. Dorkeld, F. Rechenmann, & C. Gautier
[19] "Petri Net Representations in Metabolic Pathways"
V.N. Reddy, M.L. Mavrovouniotis, & M.L. Liebman
[20] "Minimizing Complexity in Cellular Automata Models of
Self-Replication" J.A. Reggia, H.-H. Chou, S.L. Armentrout,
& Y. Peng
[21] "Building Large Knowledge Bases in Molecular Biology"
O. Schmeltzer, C. Mdigue, P. Uvietta, F. Rechenmann,
F. Dorkeld, G. Perriere, & C. Gautier
[22] "A Service-Oriented Information Sources Database for the
Biological Sciences" G.K. Springer & T.B. Patrick
[23] "Hidden Markov Models and Iterative Aligners: Study of their
Equivalence and Possibilities" H. Tanaka, K. Asai, & M. Ishikawa
[24] "Protein Structure Prediction System Based on Artificial
Neural Networks" J. Vanhala & K. Kaski
[25] "Transmembrane Segment Prediction from Protein Sequence
Data" S.M. Weiss, D.M. Cohen & N. Indurkhya
TUTORIAL PROGRAM
Tutorials will be conducted at the Bethesda Ramada Hotel on
Tuesday, July 6.
12:00-2:45pm "Introduction to Molecular Biology for Computer
Scientists" Prof. Mick Noordewier (Rutgers University)
This overview of the essential facts of molecular biology is
intended as an introduction to the field for computer scientists
who wish to apply their tools to this rich and complex domain.
Material covered will include structural and informational
molecules, the basic organization of the cell and of genetic
material, the "central dogma" of gene expression, and selected
other topics in the area of structure, function, and regulation as
relates to current computational approaches. Dr. Noordewier has
appointments in both Computer Science and Biology at Rutgers, and
has extensive experience in basic biological research in addition
to his current work in computational biology.
12:00-2:45pm "Introduction to Artificial Intelligence for
Biologists" Dr. Richard Lathrop (MIT & Arris Corp.)
An overview of the field of artificial intelligence will be
presented, as it relates to actual and potential biological
applications. Fundamental techniques, symbolic programming
languages, and notions of search will be discussed, as well as
selected topics in somewhat greater detail, such as knowledge
representation, inference, and machine learning. The intended
audience includes biologists with some computational background,
but no extensive exposure to artificial intelligence. Dr.
Lathrop, co-developer of ARIADNE and related technologies, has
worked in the area of artificial intelligence applied to
biological problems in both academia and industry.
3:00-5:45pm "Neural Networks, Statistics, and Information Theory
in Biological Sequence Analysis" Dr. Alan Lapedes (Los Alamos
National Laboratory)
This tutorial will cover the most rapidly-expanding facet of
intelligent systems for molecular biology, that of machine
learning techniques applied to sequence analysis. Closely
interrelated topics to be addressed include the use of artifical
neural networks to elicit both specific signals and general
characteristics of sequences, and the relationship of such
approaches to statistical techniques and information-theoretic
views of sequence data. Dr. Lapedes, of the Theoretical
Division at Los Alamos, has long been a leader in the use of such
techniques in this domain.
3:00-5:45pm "Genetic Algorithms and Genetic Programming"
Prof. John Koza (Stanford University)
The genetic algorithm, an increasingly popular approach to highly
non-linear multi-dimensional optimization problems, was originally
inspired by a biological metaphor. This tutorial will cover both
the biological motivations, and the actual implementation and
characteristics of the algorithm. Genetic Programming, an
extension well-suited to problems where the discovery of the size
and shape of the solution is a major part of the problem, will
also be addressed. Particular attention will be paid to
biological applications, and to identifying resources and software
that will permit attendees to begin using the methods. Dr. Koza,
a Consulting Professor of Computer Science at Stanford, has taught
this subject since 1988 and is the author of a standard text in
the field.
3:00-5:45pm "Linguistic Methods in Sequence Analysis"
Prof. David Searls (University of Pennsylvania)
& Shmuel Pietrokovski (Weizmann Institute)
Approaches to sequence analysis based on linguistic methodologies
are increasingly in evidence. These involve the adaptation of
tools and techniques from computational linguistics for syntactic
pattern recognition and gene prediction, the classification of
genetic structures and phenomena using formal language theory, the
identification of significant vocabularies and overlapping codes
in sequence data, and sequence comparison reflecting taxonomic and
functional relatedness. Dr. Searls, who holds research faculty
appointments in both Genetics and Computer Science at Penn,
represents the branch of this field that considers higher-order
syntactic approaches to sequence data, while Shmuel Pietrokovski
has studied and published with Prof. Edward Trifinov in the area
of word-based analyses.
REGISTRATION FORM
Mail, with check made out to "ISMB-93", to:
ISMB Conference, c/o J. Shavlik
Computer Sciences Department
University of Wisconsin
1210 West Dayton Street
Madison, WI 53706 USA
================================================
Name____________________________________________
Affiliation_____________________________________
Address_________________________________________
________________________________________________
________________________________________________
________________________________________________
Phone___________________________________________
FAX_____________________________________________
Electronic Mail_________________________________
Registration Status: ____ Regular ____ Student
Presenting? ____ Talk ____ Poster
================================================
TUTORIAL REGISTRATION
____"Molecular Biology for Computer Scientists"
or
____"Artificial Intelligence for Biologists"
- - - - - - - - - - - - - - - -
____"Neural Networks, Statistics, and
or Information Theory in Sequence Analysis"
____"Genetic Algorithms and Genetic Programming"
or
____"Linguistic Methods in Sequence Analysis"
================================================
PAYMENT (Early Registration Before June 1)
Registration: Early Late $___________
Regular $100 $125
Student $75 $100
Tutorials: One Two $___________
Regular $50 $65
Student $25 $35
Total: $___________
================================================
Registration fees include conference proceedings,
refreshments, and general program expenses.
ORGANIZING COMMITTEE
Lawrence Hunter NLM
David Searls U. of Pennsylvania
Jude Shavlik U. of Wisconsin
PROGRAM COMMITTEE
Douglas Brutlag Stanford U.
Bruce Buchanan U. of Pittsburgh
Christian Burks Los Alamos National Lab
Fred Cohen U.C.-San Francisco
Chris Fields Inst. for Genome Research
Michael Gribskov U.C.-San Diego
Peter Karp SRI International
Toni Kazic Washington U.
Alan Lapedes Los Alamos National Lab
Richard Lathrop MIT & Arris Corp.
Charles Lawrence Baylor
Michael Mavrovouniotis U. of Maryland
George Michaels NIH
Harold Morowitz George Mason U.
Katsumi Nitta ICOT
Mick Noordewier Rutgers U.
Ross Overbeek Argonne National Lab
Chris Rawlings ICRF
Derek Sleeman U. of Aberdeen
David States Washington U.
Gary Stormo U. of Colorado
Ed Uberbacher Oak Ridge National Lab
David Waltz Thinking Machines Corp.
------------------------------
End of ML-LIST (Digest format)
****************************************
