Co-sponsored by DIMACS and the IEEE Information Theory Society.
We see considerable overlap in research interests between information
theory and algorithm design in the area of coding problems for data
compression, with particular focus on lossless tree structured codes such
as Huffmann codes and its variants. Unfortunately, there is not as much
overlap in our perspectives. Information theorists are more concerned with
entropy-based performance bounds and tradeoffs with respect to
additional criteria of interest in the context of data compression systems,
e.g. synchronization or buffer management. Computer scientists are more
concerned
with questions of efficiency and algorithm design. These very different
perspectives on similar problems has led to a lack of
communication, occasionally resulting in situations in which problems
considered difficult by one community are actually solvable
using tools well developed and understood by the other.
We thus think a small workshop on codes and trees discussing
both the algorthmic and information theoretic approaches would be
useful in promoting the exchange of ideas between the two communities. We
are planning a small 3 day workshop, October 5 - 7, 1998 at the DIMACS
Center, with support from DIMACS (including limited financial support for
participants). We plan to publish an edited version of the workshop proceedings in the DIMACS-AMS book series.
Possible topics covered at the workshop could include but are not limited
to:
Next: Call for Participation
Workshop Index
DIMACS Homepage