See all upcoming seminars in LäsIT and seminar web pages at the homepage for the PhD studentseminars, TDB, Vi2, Theory and Applications Seminars (TAS) @ UpMARC., Department of Mathematics and The Stockholm Logic Seminar.

Disputation | Dissertation
26 August
Ricardo Alves: Leveraging Existing Microarchitectural Structures to Improve First-Level Caching Efficiency
Location: Sal VIII, Universitetshuset, Time: 9:00

Ricardo Alves will present and defend his PhD thesis Leveraging Existing Microarchitectural Structures to Improve First-Level Caching Efficiency.
Opponent: Professor Mattan Erez, Department of Electrical & Computer Engineering at The University of Texas at Austin (UTA). .
Supervisors: David Black-Schaffer and Stefanos Kaxiras.

DiVA includes an abstract and the full text of the thesis.

3 September
Adrian Wills, University of Newcastle: Maximum-Likelihood parameter estimation for Jump Markov Linear Systems
Location: ITC 2345, Time: 13:15

In this seminar we will present recent work on maximum-likelihood estimation of parameter values for Jump-Markov Linear Systems. These systems are a class of time-varying models where, at each time instant, the dynamics are selected from a finite collection of linear systems according to a so-called index variable. This index variable is allowed to randomly switch (or jump) values from one time interval to the next, which allows for rapid changes in dynamic response such as observed under fault conditions. The inherent difficulty in employing a maximum-likelihood approach for these systems is that the index variable is not measured directly, so the likelihood evaluation involves a sum over all the possible index sequences, which grows exponentially large as the number of data points increases. Therefore, existing approaches have employed various approximations to mitigate the intractable likelihood evaluation. One promising approach employs an Expectation-Maximisation method, where the expectation step is approximated using Sequential Monte Carlo techniques. In this seminar, we will similarly employ the Expectation-Maximisation method, but we will present an alternative approximation to the expectation step by employing linear filtering and smoothing theory. The new approach will be demonstrated on several simulation examples against state-of-the-art alternatives.

IT20 Talk
11 September
: Effective computations - a driving force for IT-development
Location: Polacksbacken, Hörsalen, House 6, Time: 13:15-14:15

See also the list of all upcoming seminars.

Internal seminars. Lecturers may be either internal or external.