TUTORIAL CALLS FOR PARTICIPATION
Twelfth International Conference on Machine Learning
Tahoe City, California, U.S.A.
July 9, 1995
AMORTIZED ANALYSIS FOR ON-LINE LEARNING ALGORITHMS
A new family of algorithms have been recently developed within the
Computational Learning Theory community. These algorithm have
radically different behavior from the standard algorithms that are
based on various heuristics such as gradient descent. The goal of this
tutorial is to familiarize people with this new family of algorithms,
explain the differences between the new family and existing
algorithms, and to demonstrate the practical importance of the new
family. Presented by Manfred K. Warmuth (UC Santa Cruz) and Rob
Schapire (Bell Labs).
Registration deadline: June 20, 1995.
PROBABILISTIC METHODS FOR DATA ANALYSIS
This tutorial will cover probabilistic principles and algorithms
from a cross section of fields involved in data analysis. Four
presenters will present their own perspective on data analysis. David
Heckerman (Microsoft Research) will discuss probabilistic graphical
models; Wray Buntine (Research Institute for Advanced Computing
Science, NASA Ames Research Center) will discuss software and
applications using probabilistic principles; David Stork (Machine
Learning and Perception Group, Ricoh California Research Center) will
discuss pattern classification, neural networks and machine learning;
and Jerome Friedman (Department of Statistics and Stanford Linear
Accelerator Center, Stanford University) will discuss computational
learning and statistical prediction.
Registration deadline: June 12, 1995.
For further information, please consult the conference's World-Wide