TALK: Wednesday, Jan 28, h. 11:00 - Maurizio Fiasché, On the use of Quantum inspired Spiking Neural Networks and on their integration with chaotic measurements on EEG data for modelling epileptic brain
by Announcements of talks@IDSIA
Wednesday, Jan 28, h. 11:00 at IDSIA
Maurizio Fiasché, Politecnico di Milano (POLIMI)
On the use of Quantum inspired Spiking Neural Networks and on their
integration with chaotic measurements on EEG data for modelling
epileptic brain
Abstract.
In the last 20 years a lot of works in literature analysed and proposed
several methods capable to predict the occurrence of seizures from the
electroencephalogram (EEG) of epilepsy patients. In spite of promising
results presented, more recent evaluations could not reproduce these
optimistic findings. This evaluation poses again the issue of the EEG of
epileptic patients used for the prediction of seizures used in a joint
way with other data (e.g. fMRI). Spatio- and spectro-temporal brain data
(STBD) are the most commonly collected data for measuring brain response
to external stimuli or to internal events such as epileptic seizures.
Spiking neural networks (SNN) are brain-like connectionist methods,
where the output activation function is represented as a train of spikes
and not as a potential. This and other reasons do SNN models more
biologically close to brain principles than any of the previous
Artificial Neural Networks (ANN) methods, and for this reason, are
presented as the third generation of ANN. In fact, they have great
potential for solving complicated time-dependent pattern recognition
problems defined by time series and STBD because of their inherent
dynamical representation. Nevertheless several challenges have been
reported in works in literature for SNN techniques studied. In this
presentation we want to present a particular type of SNN, the evolving
SNN (eSNN) with a Quantum inspired evolutionary technique for its
weights, parameters and features optimization. In fact, the first
challenge approached in in this presentation is the feature and
parameter optimization in a SNN. A QiEA is a novel type of
quantum-inspired evolutionary algorithm (QiEA) for feature and parameter
optimization used here in a SNN. The QiEA is inspired by the multiple
universes principle of quantum computing, such as a quantum bit and
superposition of states. QiEA represents the most recent advance in the
field of evolutionary computation thus, as second objective, the
architecture of a QiESNN is presented here and proposed for pattern
recognition on temporal brain data.
The third objective is about the presentation of a framework where
Chaotic measurement and the QiESNN presented before are used in an
integrated way for better understanding the epileptic brain. This
framework concept, architecture and applications are approached with
several possible future evolutions.
Maurizio Fiasché is research coordinator at the Department of Economics,
Management and Industrial Engineering of Politecnico di Milano (POLIMI),
where his research interests are in Computational Intelligence and ICT
for manufacturing. He received a Ph.D. in Computer Engineering in 2010
and a M.Sc. in Electronical Engineering with magna cum laude in 2006 at
University of Reggio Calabria, Italy. Member of INNS and Senior Member
of IEEE in: Computational Intelligence, EMB, Signal Processing and
Computer societies, He was involved in two IEEE WGs as member for his
expertise: IEEE P23026^(TM) Standard for Systems and software
engineering -- Engineering and management of websites for systems,
software, and services information, and IEEE P2145-1^(TM) : Standard
for Smart Transducer Interface for Sensors and Actuators - Common
Network Services. He won a best paper award during ICONIP 2008
conference in Auckland, NZ. He is author of 37 papers in international
journals and conference proceedings, member in Technical Program
committee in about 50 International Conferences, and referee for top
international Journal also including Elsevier Neural Networks, Neural
Computing and IEEE Transactions. He has been involved as Senior
Researcher in Social&Smart EU FP7 (FIRE Project) with University of
Milan, in white'R FP7 Project, and in Touchplant Project of Regione
Lombardia with Politecnico di Milano. Moreover he has been lecture for
several Universities and Institutes in Electrotechnical, Statistical
Inference, Computer Science and Signal Processing courses. He is also
senior consultant as ICT Project Manager and Software Engineer for
several national and multinational Companies and consulting Groups since
14 years.
9 years, 11 months
Thu 15/1, 11 am, Talk by Pantelis Sopakakis: A Distributed Accelerated Dual Gradient Projection Algorithm for Large-scale Stochastic Model Predictive Control
by Announcements of talks@IDSIA
The Talk will be held at IDSIA, Meeting Room, Galleria 1, Thursday 15th of January, 11 am
A Distributed Accelerated Dual Gradient Projection Algorithm for Large-scale Stochastic Model Predictive Control
P. Sopasakis and A.K. Sampathirao,
IMT Institute for Advanced Studies Lucca, Piazza San Ponziano, 6, 55100, Lucca, Italy.
Uncertainty is ubiquitous – is it measurement noise or modelling uncertainty or any other type, it inevitably needs to be taken into account when modelling a dynamical system and, of course, when a control system needs to be designed. Large-scale uncertain systems such as power production plants and power distribution grids, drinking water networks and irrigation systems, computer networks and more have attracted a lot of attention and call for advanced control methodologies to cope with the computational difficulties that arise. In this talk, I will motivate the development of advanced numerical algorithms for large-scale systems and present a distributed variant of a dual first order method for stochastic model predictive control (MPC).
Two well-established approaches have been presented in the literature for dealing with system uncertainty: robust and stochastic control. For both cases, numerous model predictive control formulations have been proposed to accommodate state and input constraints and control the system while optimising some desired performance criterion. Although stochastic control has been shown to be less conservative and enjoy plenty favourable theoretical properties, its applicability is conditioned by its complexity. In particular, two major approaches are pursued when modelling uncertainty: it is either assumed that the uncertain parameters follow the normal distribution (or some other distribution whose cumulative inverse is known exactly), or it is assumed that the uncertainty is finite-valued and evolves as a tree of scenarios. The normality assumption is often not met while the latter approach offers greater flexibility at the cost of higher computational complexity. Therefore, computational feasibility has been a limiting factor for stochastic model predictive control algorithms; overcoming this limitation was the main motivation for the work I am going to present.
General-purpose graphics processing units (GP-GPUs) are gaining increased popularity for scientific computations. They offer unprecedented computational capabilities, which may amount up to almost 9000GFlop/s in single precision or 3000GFlop/s in double precision. Our ability to master and harness this computational power depends on the availability of algorithms able to properly decompose the problem into simple sub-problems and iterate over those in lockstep (i.e., performing the same set of operations at the same time in parallel over a different set of data).
We designed an accelerated dual gradient projection algorithm for tree-based stochastic model predictive control problems and implemented it on GPUs, which afforded a minimum speed-up of 50x (with respect to a CPU implementation). We compared the performance of the proposed algorithm with Gurobi, a widely used solver for quadratic optimisation problems, and recorded speed-ups of 1 to 3 orders of magnitude. Additionally, the accuracy of the proposed algorithm can be controlled so that less accurate solutions are obtained faster for which, however, explicit sub-optimality and infeasibility bounds are known. The proposed algorithm is well-suited for stochastic MPC problems and paves the way for its application in many fields of interest.
9 years, 11 months