All Years Seminars
[INMA] 2010-06-15 (14.00) : Interpolatory model reduction for large-scale systems
At Euler building
Speaker :
Thanos ANTOULAS (Rice University, Texas, USA)
Abstract :
Interpolatory model reduction methods have emerged as effective strategies for approximation of large-scale linear dynamical systems.
These methods produce reduced models whose transfer function interpolates the original system transfer function at selected interpolation points.
They are closely related to rational Krylov-based methods and indeed, are precisely that for single input/single output systems.
One main reason why interpolatory model reduction has become the method of choice in true large-scale settings is that this class of methods is numerically stable and well-suited to large scale computations.
The approximation theory behind this approach is similar to that of rational interpolants used to approximate meromorphic functions.
The goal of this talk is to present a survey of model reduction by interpolation.
[INMA] 2010-05-25 (14.00) : Modeling of the interaction between smooth pursuit and saccadic eye movements
At Euler building
Speaker :
Coppe, Sébastien
Abstract :
The objective of this work was to implement a model which could
illustrate the orienting eye movements of human subjects.
This model generates the two main types of orienting eye movements,
the fast movements (saccades) and the slow movements (pursuit). In the
literature, there is no model illustrating the detailed interactions
between those two types of eye movements, for example to determine the
switch between fast and slow systems.
It is important to emphasize the fact that there are important delays
(more than 100 ms) in visual sensory signals. One way to deal with
these delays is to have a good interaction between the two systems.
Another way is to use prediction to anticipate target movement. These
predictive mechanisms are also used in absence of visual stimuli, for
example during temporary blanking of the target, and need as well an
efficient interaction between fast and slow systems.
The model proposed here takes into account recent experimental results
in order to explain the behavior of a normal subject performing a
visual tracking task.
This model could facilitate the interpretation of new experimental
results, based on its theoretical predictions. It could also guide
future experimental research.
[INMA] 2010-05-18 (14.00) : Towards Leakage-Resilient Cryptographic Devices
At Euler building
Speaker :
Standaert, François-Xavier
Abstract :
Traditionally, cryptographic algorithms provide security against an adversary who has only black box
access to cryptographic devices. That is, the only thing the adversary can do is to query the
cryptographic algorithm on inputs of its choice and analyze the responses, which are always
computed according to the correct original secret information. However, such a model does not always
correspond to the realities of physical implementations. During the last decade, significant attention
has been paid to the physical security evaluation of cryptographic devices. In particular, it has been
demonstrated that actual attackers may be much more powerful than what is captured by the black box model.
For example, they can actually get a side-channel information, based on the device's physical computational
steps. As a consequence, some kind of obfuscation is required to protect integrated circuits from these
physical attacks. This is especially important for small embedded devices (e.g. smart cards, RFIDs, sensor
networks, ...) that can typically be under and adversary's control for a short period of time. This implies
new theoretical concerns (how to exactly model and evaluate these physical threats) and practical ones
(how to prevent them). In this talk, I will discuss different results in the area of side-channel attacks,
with a particular focus on formal tools that can be used to evaluate physical security on a fair basis.
Starting from an introductive view of the field, I will describe some well known attacks and countermeasures,
present a framework for the analysis of side-channel key-recovery attacks and finally discuss the connection
of this framework with recent works in leakage-resilient cryptography.
[INMA] 2010-05-11 (14.00) : Dequantizing Compressed Sensing: Recovering Quantization Consistency with Oversampling and Non-Gaussian Constraints
At Euler building
Speaker :
Jacques, Laurent
Abstract : Compressed Sensing is a recent revolution in (general) signal sampling whose advent is due to the seminal works
of E. Candès, T. Tao, J. Romberg and D. Donoho in 2006. In short, this theory shows that, without loss of information, a
signal can be sampled in function of its intrinsic dimension rather than according to its cutoff frequency. The reason of
this seemingly magic statement comes from (i) a generalization of the sampling into any linear (and often random) measurement process,
(ii) a prior knowledge of the signal structure, i.e. its assumed sparsity or compressibility in a certain basis (e.g. wavelets, DCT, curvelets, ...),
and (iii) the use of non-linear reconstruction techniques (e.g. Basis Pursuit, Lasso, Greedy methods, ...).
Quantization of measurements is however a common procedure that Compressed Sensing theory has to consider carefully. First, on a practical point
of view, any coder or device integrating a CS encoding, i.e. the random projections of the signal, has to transmit data to the decoder in a digital
way so as for instance to limit transmission errors. Second, from an Information Theory standpoint, it is better to consider the real compression
achieved by CS by counting the number of bits necessary to represent the signal with a controlled error, i.e. a binary representation achieved by
quantization.
In this talk, after a brief introduction of the CS theory, I will present a new reconstruction method, called Basis Pursuit DeQuantizer (BPDQ),
to be used when the random measurements of a signal have undergone a scalar quantization. This reconstruction relies on a convex optimization that
minimizes the sparsity of the signal to reconstruct under a fidelity constraint that is not expressed in the l2-norm anymore, as for the common Basis
Pursuit DeNoising (BPDN) reconstruction, but in the lp-norm (with p >= 2). Taking p higher than 2 provides indeed a better Quantization Consistency of
the reconstructed signal, i.e. the fact that quantizing the remeasured reconstruction is close to the original measurements. For the non-uniform case,
we show furthermore that a simple "weighting" of the lp-norm, with weight related to the quantization bin widths, is sufficient to provide provable
guarantees on the quality of the reconstructed signal.
An important message of this work is that the quality of the reconstruction improves with p in oversampled situation, i.e. when the number of measurements
is higher than a minimal value increasing with p. The reason of this stability relies on a generalization of the Restricted Isometry Property (RIP),
conditioning somehow the CS measurement matrix, to any embedding between two normed spaces. This property is again respected with a controlled
probability by Random Gaussian matrices when these two spaces are the (weighted) lp Banach space and the l2 Hilbert space, i.e. the case used to prove
BPDQ stability. We conclude this talk by providing some simulations of signal and image reconstructions from quantized random measurements.
[INMA] 2010-05-04 (14.00) : The Clouds of Revolution
At Euler building
Speaker :
Peter VAN ROY (UCL)
Abstract : Cloud computing is a recent technology that provides metered computational and storage services at multiple granularities
for a specified level of quality. It makes it possible for companies to manage their IT infrastructures more economically. But this is
only the tip of the iceberg: cloud computing is much more than that. It starts with three notable properties: (almost infinite)
scalability of resources, elasticity (the ability to scale resource usage up and down rapidly), and pay-per-use (only pay for what you use).
These properties enable a whole new class of applications that will (in our view) revolutionize the Internet yet again. For example,
we predict that in the next ten years, voice communication will do real-time language translation. Implementing this requires enormous computational
and storage requirements for short periods, and this is exactly what the cloud provides. For companies that invest in this area now it is an opportunity
to be part of a new market from its beginning.
Seminars
Subscribe (ICS)
INMA Contact Info
Mathematical Engineering (INMA)
L4.05.01
Avenue Georges Lemaître, 4
+32 10 47 80 36
secretaire-inma@uclouvain.be
Mon – Fri 9:00A.M. – 5:00P.M.
JOIN US