On Jan. 23, 1913, the first use of the Markov chain was demonstrated in a lecture in St.
Petersburg by A. A. Markov. IACS will celebrate the history, intellectual power and
21st-century applications of Markov chains with a trio of talks on Wednesday, Jan. 23.
This public event is part of our January ComputeFest. All are welcome to attend.
WHAT: 100 Years of Markov Chains, a centenary celebration
WHEN: Wednesday, Jan. 23, 9:15 am - 12:15 pm
WHERE: 209 Pierce Hall, 29 Oxford Street
PROGRAM AND SPEAKER INFORMATION:
http://computefest.seas.harvard.edu/markov
SPEAKERS:
Brian Hayes
Senior Writer, American Scientist magazine
"First Links in the Markov Chain: Poetry and Probability"
Abstract:
On January 23, 1913, the Russian mathematician A. A. Markov presented a paper at the
Imperial Academy of Sciences in St. Petersburg describing his careful enumerations of
vowels and consonants in Alexander Pushkin's poem Eugene Onegin. A century later, the
techniques that Markov discussed that day are in daily use everywhere in statistics and
scientific computing. We call them Markov chains. On the occasion of this hundredth
anniversary I want to consider what motivated Markov's work in this area, and how he
came to illustrate his mathematical ideas with an analysis of poetic language. I will go
on to discuss a few latter-day applications of Markov chains in linguistics, including the
mass production of random prose meant to foil your spam filter.
-----
Ryan Prescott Adams
Assistant Professor of Computer Science, SEAS
"From Markov to Pearl: Conditional Independence as a Driving Principle
for Probabilistic Modeling"
Abstract:
The Markov chain is one of the fundamental abstractions for consideration of stochastic
systems. The remarkable insight of Markov was that complex phenomena can be described by
the evolution of a "memoryless" system. Markov chain theory has had an enormous
impact on probabilistic computation, natural language processing, and information theory,
among many other fields. In recent decades, Judea Pearl and others recognized that this
notion of "conditional independence" could be used more broadly to define rich
classes of probability distributions for complex natural phenomena. I will give an
overview of how these ideas connect strongly with graph theory, leading to the concept of
a probabilistic graphical model, a centerpiece of modern machine learning and statistics.
-----
Pavlos Protopapas
Research Associate, Harvard-Smithsonian Center for Astrophysics; Lecturer in Computational
Science, SEAS
"Applications of Markov Chains in Science"
Abstract:
The Markov chain and its extension, Markov Chain Monte Carlo, are among the most used
algorithms in the sciences. These methods have transformed the way we do science in the
last twenty years. In the first part of this talk I will introduce the basic ideas behind
these methods and demonstrate them with simple examples: a drunk walking along a pavement,
a mutating virus, card shuffling and so on. I will then review applications of Markov
chains in various sciences, describing the latest developments and looking toward how
Markov's insight might continue to shape computational science in the future.
**********************
COMPUTEFEST REMINDERS:
SYMPOSIUM: Computing @ Exascale, the Second Annual Symposium on the Future of Computation
in Science and Engineering, Friday, Jan. 25, 9:30 am-5 pm, in Maxwell Dworkin G115.
Details:
http://computefest.seas.harvard.edu/exascale-symposium
ONLINE WORKSHOP REGISTRATION for ComputeFest has closed; walk-in registration is welcomed
as space is available 15-30 minutes before each workshop. Schedule here:
http://computefest.seas.harvard.edu/workshops
Signup for this year's STUDENT COMPUTATIONAL CHALLENGE continues through Thursday,
Jan. 17. See
http://computefest.seas.harvard.edu/student-computational-challenge. Prizes
will be presented at the Friday symposium.
-----------------
Rosalind Reid
Executive Director, Institute for Applied Computational Science
Harvard School of Engineering and Applied Sciences
Pierce Hall, 29 Oxford Street
Cambridge, MA 02138
http://iacs.seas.harvard.edu/people
rreid(a)seas.harvard.edu | 617-384-9091
_______________________________________________
Iacs-events mailing list
Iacs-events(a)seas.harvard.edu
https://lists.seas.harvard.edu/mailman/listinfo/iacs-events