Received: from ics.uci.edu by paris.ics.uci.edu id aa23675; 7 Oct 94 12:34 PDT Received: from ics.uci.edu by q2.ics.uci.edu id aa13170; 7 Oct 94 12:33 PDT To: ML-LIST: ; Subject: Machine Learning List: Vol. 6 No. 26 Reply-to: ml@ics.uci.edu Date: Fri, 07 Oct 1994 12:21:28 -0700 From: Michael Pazzani Message-ID: <9410071233.aa13170@q2.ics.uci.edu> Machine Learning List: Vol. 6 No. 26 Friday, October 7, 1994 Contents: JAIR ML paper Kolmogorov complexity, priors, algorithmic art Machine Learning course available New release of PEBLS system now available AI Faculty positions at UC Irvine Preliminary Call for Papers ML95 The AI and Statistics conference in '95 is strong on learning! AI/Stats Workshop 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/ or N.Z where X and N are the volume and number of the issue; ID: anonymous PASSWORD: URL- http://www.ics.uci.edu/AI/ML/Machine-Learning.html ---------------------------------------------------------------------- Date: Mon, 26 Sep 94 15:44:13 PDT From: Steve Minton Subject: JAIR ML paper Readers of ML-list my be interested in the following paper recently published in JAIR: Safra, S. and Tennenholtz, M. (1994) "On Planning while Learning", Volume 2, pages 111-129. PostScript: volume2/safra94a.ps (202K) Abstract: This paper introduces a framework for Planning while Learning where an agent is given a goal to achieve in an environment whose behavior is only partially known to the agent. We discuss the tractability of various plan-design processes. We show that for a large natural class of Planning while Learning systems, a plan can be presented and verified in a reasonable time. However, coming up algorithmically with a plan, even for simple classes of systems is apparently intractable. We emphasize the role of off-line plan-design processes, and show that, in most natural cases, the verification (projection) part can be carried out in an efficient algorithmic manner. The PostScript file is available via: -- comp.ai.jair.papers -- World Wide Web at http://www.cs.washington.edu/research/jair/home.html -- Anonymous FTP from either of the two sites below: CMU: p.gp.cs.cmu.edu directory: /usr/jair/pub/volume2 Genoa: ftp.mrg.dist.unige.it directory: pub/jair/pub/volume2 -- automated email. Send mail to jair@cs.cmu.edu or jair@ftp.mrg.dist.unige.it with the subject AUTORESPOND, and the body GET VOLUME2/SAFRA94A.PS (either upper or lowercase is fine). -- JAIR Gopher server: At p.gp.cs.cmu.edu, port 70. For more information about JAIR, check out our WWW or FTP sites, or send electronic mail to jair@cs.cmu.edu with the subject AUTORESPOND and the message body HELP, or contact jair-ed@ptolemy.arc.nasa.gov. ------------------------------ From: Juergen Schmidhuber Subject: Kolmogorov complexity, priors, algorithmic art Date: Fri, 30 Sep 1994 18:33:54 +0100 Wolpert writes: >> Kolmogorov complexity theory has nothing to do with the >> nfl statements. The statements hold independent of where on the >> Chomsky hierarchy the learning algorithms lie. In fact, the >> algorithms can have different computational abilities and the >> results still hold.'' My little complexity argument does not address the issue of ``where on the Chomsky hierarchy the learning algorithms lie''. It does not care. It just goes like this: Consider all finite relations between finite (bit)strings and finite (bit)strings, and all possible ways of choosing training sets and (non-overlapping) test sets. Without reference to any particular prior, a simple counting argument shows: In almost all cases, the shortest algorithm computing the test set from the training set essentially will have the size of the trivial algorithm listing the whole test set (the programming language does not matter). Therefore, in almost all cases, (1) knowledge of the training set does not tell us anything about the test set, (2) there is no hope for generalization. This is similar in spirit to what has been said in the recent discussion, and indicates that one should not be surprised by negative results concerning generalization capability in almost all cases. Admittedly, however, Wolpert is right by saying that this ``does not _directly_ address the same scenario as the nfl results''. Perhaps somebody out there is interested in working out precise formal relationships. An additional remark on the prior problem: With infinitely many (but enumerable) solution candidates, but without problem specific knowledge, it seems that we ought to be glad about a discrete enumerable prior that assigns to every solution candidate a probability at least as high as the one assigned by any other such prior P (ignoring a constant factor depending only on P). A remarkable property of the Solomonoff-Levin distribution (or universal prior) P_U is this: P_U dominates all discrete enumerable semimeasures P (including probability distributions) in the sense that for all P there is a constant c such that P_U(x) >= cP(x) for all x. ------------------------------ Date: Thu, 29 Sep 94 8:10:28 EDT From: Tom.Mitchell@cs.cmu.edu Subject: Machine Learning course available MACHINE LEARNING COURSE NOTES AVAILABLE ON MOSAIC The lecture slides and syllabus for CMU's course on Machine Learning are now available on the web. Feel free to use the slides or handouts in your own courses if you find them helpful. This material is from the fall 1994 course at CMU, offered to upper-level undergraduates and graduate students. Suggestions for improvements are solicited! Also, any good homework problems. (suggestions -> Tom.Mitchell@cmu.edu). The URL is http://www.cs.cmu.edu:8001/afs/cs.cmu.edu/usr/avrim/www/ML94/courseinfo.html Tom Mitchell and Avrim Blum ------------------------------ From: Steven Salzberg Sender: salzberg@blaze.cs.jhu.edu Date: Tue, 4 Oct 94 08:16:33 EDT Subject: new release of PEBLS system now available A new release of the PEBLS system, PEBLS 3.0, is now available via anonymous FTP. PEBLS is a nearest-neighbor learning system designed for applications where the instances have symbolic feature values. PEBLS has been applied to the prediction of protein secondary structure and to the identification of DNA promoter sequences. A technical description appears in the article by Cost and Salzberg, Machine Learning journal 10:1 (1993). PEBLS 3.0 is written entirely in ANSI C. It is thus capable of running on a wide range of platforms. Version 3.0 incorporates a number of additions to version 2.1 (released in 1993) and to the original PEBLS described in the paper: S. Cost and S. Salzberg. A Weighted Nearest Neighbor Algorithm for Learning with Symbolic Features, Machine Learning, 10:1, 57-78 (1993). PEBLS 3.0 now makes it possible to draw more comparisons between nearest-neighbor and probabilistic approaches to machine learning, by incorporating a capability for tracking statistics for Bayesian inferences. The system can thus serve to show specifically where nearest-neighbor and Bayesian methods differ. The system is also able to perform tests using simple distance metrics (overlap, Euclidean, Manhattan) for baseline comparisons. Research along these lines was described in the following paper: J. Rachlin, S. Kasif, S. Salzberg, and D. Aha. Towards a Better Understanding of Memory-Based and Bayesian Classifiers. {\it Proceedings of the Eleventh International Conference on Machine Learning} (pp. 242-250). New Brunswick, NJ, July 1994, Morgan Kaufmann Publishers. TO OBTAIN PEBLS BY ANONYMOUS FTP ________________________________ The latest version of PEBLS is available free of charge, and may be obtained via anonymous FTP from the Johns Hopkins University Computer Science Department. To obtain a copy of PEBLS, type the following commands: UNIX_prompt> ftp blaze.cs.jhu.edu [Note: the Internet address of blaze.cs.jhu.edu is 128.220.13.50] Name: anonymous Password: [enter your email address] ftp> bin ftp> cd pub/pebls ftp> get pebls.tar.Z ftp> bye [Place the file pebls.tar.Z in a convenient subdirectory.] UNIX_prompt> uncompress pebls.tar.Z UNIX_prompt> tar -xf pebls.tar [Read the files "README" and "pebls_3.doc"] For further information, contact: Prof. Steven Salzberg Department of Computer Science Johns Hopkins University Baltimore, Maryland 21218 Email: salzberg@cs.jhu.edu PEBLS 3.0 IS INTENDED FOR RESEARCH AND EDUCATIONAL PURPOSES ONLY. PEBLS 3.0 may be used, copied, and modified freely for this purpose. Any commercial or for-profit use of PEBLS 3.0 is strictly prohibited without the express written consent of Prof. Steven Salzberg, Department of Computer Science, The Johns Hopkins University. ------------------------------ Subject: AI Faculty positions at UC Irvine Date: Thu, 06 Oct 1994 16:31:35 -0700 From: Michael Pazzani UNIVERSITY OF CALIFORNIA, IRVINE Department of Information and Computer Science Faculty Positions in Artificial Intelligence The Department of Information and Computer Science (ICS) is seeking to fill a possible assistant professor and a possible associate professor position in the area of Artificial Intelligence. Research emphases of interest include, but are not limited to, automated reasoning, machine learning, neural networks, and planning. We are looking for candidates with strong research records who would thrive in a highly productive setting. Duties include undergraduate and graduate teaching in computer science. Applicants must possess a Ph.D. Candidates should show excellent promise of a distinguished research career. There are currently 4 Faculty with 25 students pursuing Ph.D.s in artificial intelligence and several international scholars working with the artificial intelligence group. The Artificial Intelligence faculty have research funding from agencies such as ARPA, AFOSR, NSF, and ONR as well as industrial partners. In addition to artificial intelligence, the ICS Department has research groups in the areas of algorithms and data structures, computer networks and distributed systems, computer systems design, educational technology, parallel processing, social and managerial analysis of computing, and software. The ICS Department is an independent campus unit reporting to the Executive Vice Chancellor. ICS faculty emphasize core computer science as well as research in emerging areas of the discipline, with effective inter-disciplinary ties to colleagues in management, neurobiology, cognitive science, engineering, and the social sciences. The department currently has 25 full-time faculty and 130 Ph.D. students. Graduate student, research, and administrative computing equipment includes 90 Macintoshes, a Sequent multiprocessor, more than 200 Sun workstations (Sparc 1s and 2s, Sun-3s and Sun-4s), more than 20 fileservers, a MasPar, and an assortment of PCs. Departmental undergraduate instructional computing equipment consists of 150 Macintoshes, a Sequent multiprocessor, 25 Sun workstations, and a large SPARC fileserver. All our major workstations and computers are tied together with networks, which are gatewayed to the campus network, and from there to the Internet. In addition, department members have access to campus-wide computing resources as well as regional super-computer access. UC-Irvine is located in Orange County, three miles from the Pacific Ocean near Newport Beach, and approximately forty miles south of Los Angeles. The campus is situated in the heart of a national center of high-technology enterprise. Both the campus and the enterprise area offer exciting professional and cultural opportunities. Salaries and benefits are competitive. Mortgage and housing assistance are available. Housing options include newly built, for-sale housing located on campus and within short walking distance from the Department. Send resume and contact information for four references to: Artificial Intelligence Position Lisa Tellier Department of Information and Computer Science University of California, Irvine Irvine, CA 92717-3425 Application screening will begin immediately upon receipt of curriculum vitae. Maximum consideration will be given to applications received by December 15, 1994. The University of California is an Affirmative Action/Equal Opportunity Employer, committed to excellence through diversity. ------------------------------ Date: Mon, 3 Oct 1994 16:29:39 -0700 From: "Jeffrey C. Schlimmer" Subject: Preliminary Call for Papers ML95 PRELIMINARY CALL FOR PAPERS Twelfth International Conference on Machine Learning Tahoe City, California July 9-12, 1995 The Twelfth International Conference on Machine Learning (ML95) will be held at the Granlibakken Resort in Tahoe City, California during July 9-12, 1995, with informal workshops on July 9. We invite paper submissions from researchers in all areas of machine learning. The conference will include presentations of refereed papers and invited talks. REVIEW CRITERIA Each submitted paper will be reviewed by at least two members of the program committee and will be judged on significance, originality, and clarity. Papers submitted simultaneously to other conferences must clearly state so on the title page. PAPER FORMAT Submissions must be clearly legible, with good quality print. Papers are limited to a total of twelve (12) pages, EXCLUDING title page and bibliography, but INCLUDING all tables and figures. Papers must be printed on 8-1/2 x 11 inch paper or A4 paper using 12 point type (10 characters per inch) with no more than 38 lines per page and 75 characters per line (e.g., LaTeX 12 point article style). The title page must include an abstract and email and postal addresses of all authors. Papers without this format will not be reviewed. To save paper and postage costs please use DOUBLE-SIDED printing. REQUIREMENTS FOR SUBMISSION Send four (4) copies of each submitted paper to one of the conference co-chairs. Papers must be received by FEBRUARY 7, 1995 . Electronic or FAX submissions are not acceptable. Notification of acceptance or rejection will be mailed to the first (or designated) author by March 22, 1995. Camera-ready accepted papers are due on April 25, 1995. INFORMAL WORKSHOPS Proposals for informal workshops are invited in all areas of machine learning. Send a two (2) page description of the proposed workshop, its objectives, organizer(s), and expected number of attendees to the workshop chair. Proposals must be received by DECEMBER 1, 1994. Conference Co-Chairs Armand Prieditis Department of Computer Science University of California Davis, CA 95616 priediti@cs.ucdavis.edu Stuart Russell Computer Science Division University of California Berkeley, CA 94720 russell@cs.berkeley.edu Program Committee (To Be Announced). Workshop Chair Sridhar Mahadevan Department of Computer Science and Engineering University of Southern Florida 4202 East Fowler Avenue, EBG 118 Tampa, Florida 33620 mahadeva@csee.usf.edu Publicity Chair Jeff Schlimmer School of Electrical Engineering and Computer Science Washington State University Pullman, WA 99164-2752 schlimme@eecs.wsu.edu http://www.eecs.wsu.edu/~schlimme Local Arrangements Debbie Chadwick Department of Computer Science University of California Davis, CA 95616 chadwick@cs.ucdavis.edu GENERAL INQUIRIES Please send general inquiries to ml95@cs.ucdavis.edu . To receive future conference announcements please send a note to the publicity chair. Current conference information available online on the World-Wide Web as http://www.eecs.wsu.edu/~schlimme/ml95.html . Jeffrey C. Schlimmer, Asst. Prof., School of EE & CS, Washington State University, Pullman, WA 99164-2752, (509) 335-2399, (509) 335-3818 FAX ------------------------------ Date: Sun, 25 Sep 94 13:14:15 PDT From: Wray Buntine Subject: the AI and Statistics conference in '95 is strong on learning! Just looking at the program of the AI and Statistics in Florida, Jan. '95. This years primary theme is "learning from data". Not only is there a strong contingent of learning papers, there is also a very impressive selection of tutorials on learning, or in areas related to learning: Machine Learning (Aha) Statistical Methods for Inducing Models from Data (Steffen Lauritzen) Probabilistic Models of Causality (Glenn Shafer) Statistical Models for Function Estimation and Classification (Trevor Hastie) Steffen Lauritzen and Trevor Hastie largely cover different areas in statistics, so if you're interested in a statistical view of machine learning, I'd recommend attending both. So if you interested in finding out about the interface between machine learning and statistics, I'd highly recommend AI and Statistics in '95. Dealine for early (cheap) regististration is 1st December. ------------------------------ Date: Sat, 1 Oct 1994 10:40:52 +0600 From: "Douglas H. Fisher" Subject: AI/Stats Workshop Preliminary Call for Participation Fifth International Workshop on ARTIFICIAL INTELLIGENCE and STATISTICS January 4-7, 1995 Ft. Lauderdale, Florida TECHNICAL and TUTORIAL PROGRAM: This is the fifth in a series of workshops that has brought together researchers in Artificial Intelligence and in Statistics to discuss problems of mutual interest. To encourage interaction and a broad exchange of ideas, there will be 20 discussion papers in single session meetings over three days (Jan. 5-7). Two poster sessions will provide the means for presenting and discussing the remaining research papers. Attendance at the workshop is *not* limited to paper presenters. The three days of research presentations will be preceded by a day of tutorials (Jan. 4). The tutorial topics, presenters, and approximate times are: (1) Machine Learning 9:00AM - 12:15PM (Dr. David Aha, Naval Research Lab) (2) Statistical Methods for Inducing 9:00AM - 12:15PM Models from Data (Prof. Steffen Lauritzen, Aalborg U.) (3) Probabilistic Models of Causality 2:00PM - 5:15PM (Prof. Glenn Shafer, Rutgers U.) (4) Statistical Models for Function 2:00PM - 5:15PM Estimation and Classification (Prof. Trevor Hastie, Stanford U.) Notes prepared by the tutorial presenters will be made available at the Workshop. LOCATION: The 1995 Workshop will be held at Pier Sixty Six Resort & Marina 2301 SE 17th Street Causeway Fort Lauderdale, Florida, 33316 USA. Phone: 800-327-3796 (outside Florida) 305-525-6666 Fax : 305-728-3541 The hotel is a 22 acre resort located on the intracoastal waterway. Available amenities include two pools, a 40 person hydrotherapy pool, spa, tennis courts, a children's activity club, seven restaurants and lounges, and water shuttle service to the beach. The Hotel is most conveniently reached from Fort Lauderdale International Airport, which is about 5-10 minutes by car/cab. The Hotel is approximately 45-60 minutes by car from Miami International Airport. The Resort is holding a block of rooms at the rate of $95 US dollars (for single/double) until Dec. 10, 1994. Reservations should be made before this date. The block is held under the name `SOCIETY for ARTIficial Intelligence and Statistics' (or SOCIETY ARTI). REGISTRATION: Registration for the Technical Program (plenary and poster sessions) includes a proceedings of papers submitted by authors, continental breakfasts each day of the technical program, and tentatively, two lunches and one dinner. The Workshop offers student rates and an early-registration discount. Registration rates and instructions can be found on the Registration Form at the end of this Call. Registration for tutorials can also be made in advance using the Registration Form. PROGRAM COMMITTEE: General Chair: D. Fisher Vanderbilt U., USA Program Chair: H. Lenz Free U. Berlin, Germany Members: W. Buntine NASA (Ames), USA J. Catlett AT&T Bell Labs, USA P. Cheeseman NASA (Ames), USA P. Cohen U. of Mass., USA D. Draper U. of Bath, UK Wm. Dumouchel Columbia U., USA A. Gammerman U. of London, UK D. J. Hand Open U., UK P. Hietala U. Tampere, Finland R. Kruse TU Braunschweig, Germany S. Lauritzen Aalborg U., Denmark W. Oldford U. of Waterloo, Canada J. Pearl UCLA, USA D. Pregibon AT&T Bell Labs, USA E. Roedel Humboldt U., Germany G. Shafer Rutgers U., USA P. Smyth JPL, USA Tutorial Chair: P. Shenoy U. Kansas, USA MORE INFORMATION: For more information write dfisher@vuse.vanderbilt.edu or call 615-343-4111. SPONSORS: Society for Artificial Intelligence and Statistics International Association for Statistical Computing *********** Papers accepted for Technical Program Fifth International Workshop on Artificial Intelligence and Statistics PLENARY PAPERS Almond, Schimert (MathSoft) Missing data models as meta-data Brent, Murthy, Lundberg Minimum description length induction (John Hopkins U) for discovering morphemic suffixes Buntine (NASA Ames) Software for data analysis with graphical models: basic tools Chickering, Geiger, Heckerman Learning Bayesian networks: search (MicroSoft) methods and experimental results Cohen, Gregory, Ballesteros, Two algorithms for inducing structural St Amant (U Mass) equation models from data Cooper (U Pitt) Causal discovery from observational data in the presence of selection bias Cox (US West) Using causal knowledge to learn more useful decision rules from data Decatur (Harvard U) Learning in hybrid noise environments using statistical queries Elder (Rice U) Heuristic search for model structure Gebhardt, Kruse Learning possibilistic networks from data (U Braunschweig) Kasahara, Ishikawa, Viewpoint-based measurement of semantic Matsuzawa, Kawaoka similarity between words (Nippon TT) Lubinsky (U Witwatersrand SA) Structured interpretable regression Madigan, Almond (U Washington) Test selection strategies for belief networks Malvestuto (U L'Aquila, IT) Derivation DAGs for inferring interaction models Merz (U Cal Irvine) Dynamic learning bias selection Pearl (UCLA) A causal calculus for statistical research with applications to observational and experimental studies Riddle, Frenedo, Newman Framework for a generic knowledge (Boeing) discovery tool Shafer, Kogan, Spirtes A generalization of the Tetrad (Rutgers) representation theorem St Amant, Cohen (U Mass) Preliminary design for an EDA assistant Yao, Tritchler (U Toronto) Likelihood-based causal inference POSTER PAPERS Aha, Bankert (NRL) A comparative evaluation of sequential feature selection algorithms Ali, Brunk, Pazzani Learning multiple relational rule-based (U Cal Irvine) models Almond (MathSoft) Hypergraph grammars for knowledge-based model construction Anderson, Carlson, Westbrook Tools for analyzing AI programs Hart, Cohen (U Mass) Bergman, Rivest (MIT) Picking the best expert from a sequence Blau (U Rochester) Ploxoma: Test-bed for uncertain inference Breese, Heckerman Probabilistic case-based reasoning (MicroSoft) Burke (U Nevada) Comparing the prediction accuracy of statistical models and artificial neural networks in breast cancer Catlett (ATT) Tailoring rulesets to misclassification cost Chen, Yeh Predicting stock returns with genetic (National Chengchi U) programming Cheng (U Cincinnati) Analysis and Application of the Generalized Mean-Shift Process Cozman, Krotkov (CMU) Truncated Gaussians as tolerance sets Cunningham (U Waikato) Textual data mining De Vel, Li, Coomans Non-Linear dimensionality reduction: (U James Cook, NZ) A comparative performance study DuMouchel, Friedman, Johnson Natural language processing of Hripcsak (Columbia U) radiology reports Esposito, Malerba, Semeraro A further study of pruning methods in (U degli Studi, IT) decision tree induction Feelders, Verkooijen Which method learns most from the data? (U Twente, Netherlands) Franz (CMU) Classifying new words for robust parsing Gelsema (Erasmus U, Abductive reasoning in Bayesian belief The Netherlands) networks using a genetic algorithm Harner, Galfalvy Omega-Stat: An environment for (West Virginia U) implementing intelligent modeling strategies Heckerman, Shachter A decision-based view of causality (MicroSoft) Howe (Colorado St U) Finding dependencies in event streams using local search Jenzarli (U Tampa) Solving influence diagrams using Gibbs sampling John (Stanford U) Robust linear discriminant trees Ketterlin, Gancarski, Korczak Hierarchical clustering of composite (U Louis Pasteur) objects with a variable number of components Kim (Korea Adv. Inst. of Sci. An approach to fitting large influence and Eng.) diagrams Kim, Moon (Syracuse U) Modeling life time data by neural networks Kloesgen (German Nat. Rsch.) Learning from data: Pattern evaluations and search strategies Larranaga, Murga, Poza, Structure learning of Bayesian networks Kuijpers (U Basque, by hybrid genetic algorithms Spain) Lekuona, Lacruz, Lasala Graphical models for dynamic systems (U de Zaragoza, Spain) Liu (U Kansas) Propagation of Gaussian belief functions Martin (U Cal, Irvine) A hypergeometric null hypothesis probability test for feature selection and stopping Martin (U Cal, Irvine) Evaluating and comparing classifiers: Complexity measures Murthy (John Hopkins U) Statistical preprocessing of decision trees Neufeld, Adams, Choy, Philip, Part-of-speech tagging from small Tawfik (U Saskatchewan) data sets Oates, Gregory, Cohen (U Mass) Detecting complex dependencies in categorical data Pazzani (U Cal Irvine) Searching for attribute dependencies in Bayesian classifiers Provan, Singh (Inst. for Learning ``Predictively-Optimal'' Decision Systems Res.) Bayesian Networks Risius, Seidelmann Combining statistics and AI in the (Hahn-Meitner Inst) optimization of semiconductor films for solar cells Shenoy (U Kansas) Representing and solving asymmetric decision problems using valuation networks Srkantan, Srihari Data representations in learning (SUNY Buffalo) Sun, Qiu, Cox (US West) A hill-climbing approach to construct near optimal decision trees Valtorta (U South Carolina) MENTOR: A Bayesian model for prediction and intervention in mental retardation Young, Lubinsky (UNC) Learning from data by guiding the analyst: On the representation, use, and creation of visual statistical strategies *********** Registration Form Fifth International Workshop on Artificial Intelligence and Statistics Participants may register on site. To register in advance of the Workshop send this form and a check (in US dollars) made to the order of **Society for Artificial Intelligence and Statistics** in the appropriate amount to: Doug Fisher Department of Computer Science Box 1679, Station B Vanderbilt University Nashville, Tennessee 37235 USA Advance registration discounts apply if registration is received by Dec. 1, 1994. Name: ________________________________________ Affiliation: _________________________________ Phone: _______________________________________ Fax: _________________________________________ Email: _______________________________________ Address: _____________________________________ _____________________________________ _____________________________________ Technical Program -- check one: ____ Technical Program (regular, by Dec. 1, 1994): $245 ____ Technical Program (student, by Dec. 1, 1994): $155 ____ Technical Program (regular, after Dec. 1, 1994): $295 ____ Technical Program (student, after Dec. 1, 1994): $195 Technical Program Subtotal: $____ Tutorial Program -- check applicable tutorials, if any. Note that the tutorial times may conflict; to avoid conflict at most one selection from (1) and (2), and one selection from (3) and (4) may be made. ____ (1) Machine Learning ____ (regular, by Dec. 1): $ 70 ____ (student, by Dec. 1): $ 45 ____ (regular, after Dec. 1): $ 80 ____ (student, after Dec. 1): $ 55 ____ (2) Statistical Methods for Inducing Models from Data ____ (regular, by Dec. 1): $ 70 ____ (student, by Dec. 1): $ 45 ____ (regular, after Dec. 1): $ 80 ____ (student, after Dec. 1): $ 55 ____ (3) Probabilistic Models of Causality ____ (regular, by Dec. 1): $ 70 ____ (student, by Dec. 1): $ 45 ____ (regular, after Dec. 1): $ 80 ____ (student, after Dec. 1): $ 55 ____ (4) Statistical Models for Function Estimation and Classification ____ (regular, by Dec. 1): $ 70 ____ (student, by Dec. 1): $ 45 ____ (regular, after Dec. 1): $ 80 ____ (student, after Dec. 1): $ 55 Tutorial Program Subtotal: $____ Technical and Tutorial Total: $____ ------------------------------ End of ML-LIST (Digest format) ****************************************