Copy Link
Add to Bookmark
Report
Machine Learning List Vol. 5 No. 05
Machine Learning List: Vol. 5 No. 5
Wednesday, March 10, 1993
Contents:
Machine Learning 10:1
Machine Learning 10:2
UCI Machine Learning Repository Update
Senior Research Associate: University of Sydney, Australia
Employment: NYNEX Science and Technology
ML93 Workshop on Reinforcement Learning -- Call for participation
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: Wed, 3 Mar 93 11:00:24 PST
From: Tom Dietterich <tgd@chert.CS.ORST.EDU>
Subject: Machine Learning 10:1
Machine Learning
January 1993, Volume 10, Number 1
Synthesis of UNIX Programs Using Derivational Analogy
S. Bhansali and M. T. Harandi
A Weighted Nearest Neighbor Algorithm for Learning with Symbolic
Features
S. Cost and S. Salzberg
Induction Over the Unexplained: Using Overly-General Domain Theories
to Aid Concept Learning
R. J. Mooney.
Subscriptions - Volumes 9-10-11-12 (12 issues) includes postage and
handling.
$240.00 Individual
132.00 Member AAAI, CSCSI
608.00 Institutional
Kluwer Academic Publishers
P.O. Box 358
Accord Station
Hingham, MA 02018-0358 USA
or
Kluwer Academic Publishers Group
P.O. Box 322
3300 AH Dordrecht
THE NETHERLANDS
(AAAI members please include membership number)
------------------------------
From: Tom Dietterich <tgd@chert.CS.ORST.EDU>
Date: Wed, 3 Mar 93 11:01:56 PST
Subject: Machine Learning 10:2
Machine Learning
February 1993, Volume 10, Number 2
Information Filtering: Selection Mechanisms in Learning Systems
S. Markovitch and P. D. Scott
Overfitting Avoidance as Bias:
C. Shaffer
Book Review: Creating a Memory of Causal Relationships by M. Pazzani
W. Cohen
A Reply to Cohen's Book Review of Creating a Memory of Causal Relationships
M. Pazzani
------------------------------
Subject: UCI Machine Learning Repository Update
Date: Thu, 04 Mar 93 23:48:19 -0800
From: "Patrick M. Murphy" <pmurphy@focl.ICS.UCI.EDU>
The following is a list of recently added and recently
documented (moved from the undocumented/ directory)
databases and domain theories.
Any comments or donations would be greatly appreciated.
Patrick M. Murphy (Site Librarian)
David W. Aha (Off-Site Assistant)
1. MONK's Problems
The MONK's problems (3) were the basis of a first international comparison
of learning algorithms. The result of this comparison is summarized in
"The MONK's Problems - A Performance Comparison of Different Learning
algorithms" by S.B. Thrun, J. Bala, E. Bloedorn, I. Bratko, B.
Cestnik, J. Cheng, K. De Jong, S. Dzeroski, S.E. Fahlman, D. Fisher,
R. Hamann, K. Kaufman, S. Keller, I. Kononenko, J. Kreuziger, R.S.
Michalski, T. Mitchell, P. Pachowicz, Y. Reich H. Vafaie, W. Van de
Welde, W. Wenzel, J. Wnek, and J. Zhang has been published as
Technical Report CS-CMU-91-197, Carnegie Mellon University in Dec.
1991.
One significant characteristic of this comparison is that it was
performed by a collection of researchers, each of whom was an advocate
of the technique they tested (often they were the creators of the
various methods). In this sense, the results are less biased than in
comparisons performed by a single person advocating a specific
learning method, and more accurately reflect the generalization
behavior of the learning techniques as applied by knowledgeable users.
2. Student Loan Relational Database
The Student Loan relational database and domain theory has 10+
intesionally and extensionally defined relations. The target
predicate no_payment_due is true for those people not required
to repay a student loan. The domain theory is consistent with
the 1000 instances of no_payment_due.
3. Statlog Project Databases
These databases were in used in the European StatLog project, which
involves comparing the performances of machine learning, statistical,
and neural network algorithms on data sets from real-world industrial
areas including medicine, finance, image analysis, and engineering
design. Not all of the databases used in the project are available
in the repository.
Vehicle Silhouettes:
The original purpose was to find a method of distinguishing
3D objects within a 2D image by application of an ensemble of
shape feature extractors to the 2D silhouettes of the objects.
Landsat Satellite:
The database consists of the multi-spectral values
of pixels in 3x3 neighbourhoods in a satellite image,
and the classification associated with the central pixel
in each neighbourhood. The aim is to predict this
classification, given the multi-spectral values. In
the sample database, the class of a pixel is coded as
a number.
Shuttle:
The shuttle dataset contains 9 attributes all of which are
numerical. Approximately 80% of the data belongs to class 1.
Australian Credit Approval:
This file concerns credit card applications. All attribute names
and values have been changed to meaningless symbols to protect
confidentiality of the data. This database exists elsewhere in
the repository (Credit Screening Database) in a slightly different
form.
Heart Disease:
This dataset is a heart disease database similar to a database
already present in the repository (Heart Disease databases)
but in a slightly different form. This database contains 13
attributes (which have been extracted from a larger set of 75).
Image Segmentation:
This dataset is an image segmentation database similar to a database
already present in the repository (Image segmentation database)
but in a slightly different form. The instances were drawn randomly
from a database of 7 outdoor images. The images were handsegmented
to create a classification for every pixel. Each instance is a 3x3
region.
4. Artificial Character Database
This database has been artificially generated by using a first order
theory which describes the structure of ten capitol letters of the
English alphabet and a random choice theorem prover which accounts
for etherogeneity in the instances. The capitol letters represented
are the following: A, C, D, E, F, G, H, L, P, R.
Each instance is structured and is described by a set of segments (lines)
which resemble the way an automatic program would segment an image.
Included is a set of training and testing sets and a domain theory.
5. ANN Thyroid Database
Another thyroid database suited for use in training ANNs. The problem
is to determine whether a patient referred to the clinic is hypothyroid.
Therefore three classes are built: normal (not hypothyroid),
hyperfunction and subnormal functioning. 3772 training instances and
3428 testing instances.
------------------------------
From quinlan@basser.cs.su.OZ.AU Fri Feb 26 09:26:29 1993
Date: Fri, 26 Feb 1993 09:21:02 +1100
Subject: Senior Research Associate: University of Sydney, Australia
Senior Research Associate
Empirical Learning Project
University of Sydney, Australia
The Australian Research Council has provided support for a project
investigating the design and evaluation of algorithms for constructing
theories from data. Topics of particular interest are:
* first-order (relational) empirical learning
* computation-intensive symbolic learning
* learning with continuous classes
A one-year appointment to the project will be available for commencement
in September 1993. Applicants should hold a PhD in Computer Science and
should have published extensively over a five-year period in areas
directly related to one or more of the above topics. Well-developed
programming skills are essential, with experience in C preferred.
Duties will be to carry out research relevant to the project, to publish
reports and papers on the research, and to take part in discussions and
seminars associated with the project.
For further information, please contact Ross Quinlan (Phone +61 2 692 3423,
Fax +61 2 692 3838, Email quinlan@cs.su.oz.au).
Salary range will be $45,613-$48,688. Some assistance with relocation
costs for will be provided. (This position would be especially
appropriate for someone taking a 12-month sabbatical.)
Applications close April 13 1993.
------------------------------
Date: Tue, 9 Mar 93 09:46:35 EST
From: Andrea Danyluk <danyluk@nynexst.COM>
Subject: Employment: NYNEX Science and Technology
The Expert Systems Lab at NYNEX Science and Technology has an opening
for someone to join its Adaptive Expert Systems project as a Member
of Technical Staff.
We are investigating the use of machine learning techniques for development
and maintenance of expert system knowledge bases. This effort targets
applications of relevance to the operations of the New York and New England
Telephone Companies. Our current focus is diagnosis of customer-reported
telephone troubles. This work requires learning about the domain,
experimenting with a variety of machine learning techniques, software
design and development, and direct involvement with telephone company
operations. Since this is an exploratory effort, learning, evolving, and
sharing knowledge about new techniques are important parts of the job.
Desired qualifications are:
At least an MS in Computer Science or a related field.
Significant experience in and knowledge of machine learning.
Experience and maturity in conceptualization, software design, and analysis.
Strong experience in C and Lisp. Knowledge and experience with relational
databases is a plus.
Strong communication skills - both written and oral.
Ability to work in a small cohesive team.
Strong commitment, dedication, and motivation to excel.
If you feel you are qualified and interested, please send your resume to:
Andrea Danyluk
NYNEX Science and Technology, Inc.
500 Westchester Avenue
White Plains, NY 10604
e-mail: danyluk@nynexst.com
fax: 914-644-2404
You may send your resume as plain e-mail, PostScript, US Mail, or as a
fax.
NYNEX is the parent company of New York Telephone and New England Telephone.
Science and Technology is the research and development organization of
NYNEX. We are located about 30 minutes north of New York City, in
suburban Westchester County.
------------------------------
Date: Fri, 5 Mar 93 15:09:44 -0500
From: Rich Sutton <rich@gte.COM>
Subject: ML93 Workshop on Reinforcement Learning -- Call for participation
Call for Participation
"REINFORCEMENT LEARNING: What We Know, What We Need"
an Informal Workshop to follow ML93
June 30 & July 1, University of Massachusetts, Amherst
Reinforcement learning is a simple way of framing the problem of an
autonomous agent learning and interacting with the world to achieve a
goal. This has been an active area of machine learning research for the
last 5 years. The objective of this workshop is to present concisely
the current state of the art in reinforcement learning and to
identify and highlight critical open problems.
The intended audience is all learning researchers interested in
reinforcement learning; little prior knowledge of the area will be
assumed. The first half of the workshop will consist mainly of
tutorial presentations, and the second half will define and explore
outstanding problems. The entire workshop will last approximately one
and a half days. Attendance will be open to all those registered for
the main part of ML93 (June 27-29).
Program Committee: Rich Sutton (chair), Nils Nilsson, Leslie
Kaelbling, Satinder Singh, Sridhar Mahadevan, Andy Barto, Steve
Whitehead
CALL FOR PAPERS. Papers are solicited that lay out relevant
problem areas, i.e., for the second half of the workshop. Proposals
are also solicited for polished tutorial presentations on basic
topics of reinforcement learning for the first portion of the workshop.
The program has yet to be established, but will probably look something
like the following (all names are provisional):
Session 1: Introduction
The Challenge of Reinforcement Learning, by Rich Sutton
History of RL, by Harry Klopf
Q-learning
Planning and Action Models, by Long-Ji Lin
Session 2: Theory
Dynamic Programming, by Andy Barto
Convergence of Q-learning and TD(lambda), by Peter Dayan
Session 3: Applications
TD-Gammon, by Gerry Tesauro
Robotics, by Sridhar Mahadevan
Session 4: Extensions
Prioritized Sweeping, by Andrew Moore
Eligibility Traces, by Rich Sutton
Sessions 5 & 6: Open Problems
Generalization
Hidden State (short-term memory)
Hierarchical RL
Search Control
Incorporating Prior Knowledge
Exploration
...
If you are interested in attending the RL workshop, please register by
sending a note with your name, email and physical addresses, level of
interest, and a brief description of your current level of knowledge
about reinforcement learning, to:
sutton@gte.com
OR
Rich Sutton
GTE Labs, MS-44
40 Sylvan Road
Waltham, MA 02254
------------------------------
End of ML-LIST (Digest format)
****************************************
