Relation to long term objectives denoted by
[SEC] Security
[RI] Reliable Information
1)- Descriptional complexity of regular languages, [SEC,RI]
Due to novel applications in Computer Science - from bio-informatics
to computer networks and Web technologies - automata theory has
faced renewed theoretical developments. The main objectives of this
line of research are
- Study of the non injective and non complexity preserving conversions
between combinatorial representations with the same expressive power
and the descriptional complexity of generic languages operations.
- Enumerative and random generation of languages representations with
applications in characterization studies and average case analysis.
- Development of tools for assisting research on formal and
computational models; due to its combinatorial nature,
high-performance implementations are required. These tools will
include: an open source extensible high-performance software library
for the symbolic manipulation of automata, tools for automatic drawing
of automata diagrams and a modular graphical user interface for the
manipulation of automata and other models of computation.
2) individual analysis of cryptographic protocols, [SEC]
The classical analysis of cryptographic protocols assumes that
certain probability distributions are well defined and known in
advance; but that is often not the case. A completely different
approach is the "individual analysis". Only very recently has this
kind of analysis been attempted. The continuation of the research
(which began a few years ago in LIACC and elsewhere) is the main topic
of this project.
The individual analysis is based on variants of the Kolmogorov
complexity; however, security measures based on resourced unbounded
Kolmogorov complexity are usually uncomputable and defined "apart from
a constant"; this is a serious drawback for practical applications, so
that the research will include the search for appropriate computable
alternatives to Kolmogorov complexity.
3) Formal models for reversible computation; measuring the degree of
reversibility, [RI]
The Landauer principle is the ultimate lower bound for non reversible
computations. Only reversible forms of computation can in theory be almost
dissipationless. The objectives can be divided in 3 main sub-areas
- Characterization and study of very simple reversible languages; this
is the continuation of some previous work in this area.
- Use of Kolmogorov complexity (or related concepts) to measure
the degree of irreversibility of a computation.
- Compare several theoretical models for reversible computation; these
models may include reversible automata, reversible languages, some
forms of rewrite rules and the unitary transformations of quantum
mechanics.
- Search for efficient methods for restoring previous states - such as
previous configurations of a machine or previous states of a database
system).
1. Peter Blanchard, Frank Harary and Rogério Reis,
Partitions into sum-free sets.
INTEGERS: The Electronic Journal of Combinatorial
Number Theory, 6, A7, 2006.
2. Marco Almeida, Nelma Moreira, and Rogério Reis,
Enumeration and Generation with a String Automata Representation,
accepted for publication in
Theoretical Computer Science,
special issue "Selected Papers of DCFS 2006".
3. C. Costa-Santos C, J. Bernardes, P. Vitányi, and Antunes L.,
Clustering Fetal Heart Rate Tracings by Compression,
19th IEEE International Symposium on Computer-Based Medical
Systems, 2006.
4. Armando B. Matos,
Linear programs in a simple reversible language,
Theoretical Computer Science, Vol. 290 No. 3,
2003.
5. Alexandre Pinto, André Souto, A. Matos and L. Antunes,
Commitment and Authentication Systems,
Proc. International Conference on Information Theoretic Security,
Lecture Notes in Computer Science,
Springer, 2007 (to appear).
Scholarships for post-grad (MSC and PhD) students