Institute of Computer Science

Institute of Computer Science 1,617 records found  beginprevious21 - 30nextend  jump to record: Search took 0.01 seconds. 
Application of the Infinitely Many Times Repeated BNS Update and Conjugate Directions to Limited-Memory Optimization Methods
Vlček, Jan ; Lukšan, Ladislav
To improve the performance of the L-BFGS method for large scale unconstrained optimization, repeating of some BFGS updates was proposed. Since this can be time consuming, the extra updates need to be selected carefully. We show that groups of these updates can be repeated infinitely many times under some conditions, without a noticeable increase of the computational time. The limit update is a block BFGS update. It can be obtained by solving of some Lyapunov matrix equation whose order can be decreased by application of vector corrections for conjugacy. Global convergence of the proposed algorithm is established for convex and sufficiently smooth functions. Numerical results indicate the efficiency of the new method.
Absolute Value Mapping
Rohn, Jiří
We prove a necessary and sufficient condition for an absolute value mapping to be bijective. This result simultaneously gives a characterization of unique solvability of an absolute value equation for each right-hand side.
Fulltext: Download fulltextPDF
Overdetermined Absolute Value Equations
Rohn, Jiří
We consider existence, uniqueness and computation of a solution of an absolute value equation in the overdetermined case.
Fulltext: Download fulltextPDF
The scalar-valued score functions of continuous probability distribution
Fabián, Zdeněk
In this report we give theoretical basis of probability theory of continuous random variables based on scalar valued score functions. We maintain consistently the following point of view: It is not the observed value, which is to be used in probabilistic and statistical considerations, but its 'treated form', the value of the scalar-valued score function of distribution of the assumed model. Actually, the opinion that an observed value of random variable should be 'treated' with respect to underlying model is one of main ideas of the inference based on likelihood in classical statistics. However, a vector nature of Fisher score functions of classical statistics does not enable a consistent use of this point of view. Instead, various inference functions are suggested and used in solutions of various statistical problems. Inference function of this report is the scalar-valued score function of distribution.
Laplacian preconditioning of elliptic PDEs: Localization of the eigenvalues of the discretized operator
Gergelits, Tomáš ; Mardal, K.-A. ; Nielsen, B. F. ; Strakoš, Z.
This contribution represents an extension of our earlier studies on the paradigmatic example of the inverse problem of the diffusion parameter estimation from spatio-temporal measurements of fluorescent particle concentration, see [6, 1, 3, 4, 5]. More precisely, we continue to look for an optimal bleaching pattern used in FRAP (Fluorescence Recovery After Photobleaching), being the initial condition of the Fickian diffusion equation maximizing a sensitivity measure. As follows, we define an optimization problem and we show the special feature (so-called complementarity principle) of the optimal binary-valued initial conditions.
On the Optimal Initial Conditions for an Inverse Problem of Model Parameter Estimation - a Complementarity Principle
Matonoha, Ctirad ; Papáček, Š.
This contribution represents an extension of our earlier studies on the paradigmatic example of the inverse problem of the diffusion parameter estimation from spatio-temporal measurements of fluorescent particle concentration, see [6, 1, 3, 4, 5]. More precisely, we continue to look for an optimal bleaching pattern used in FRAP (Fluorescence Recovery After Photobleaching), being the initial condition of the Fickian diffusion equation maximizing a sensitivity measure. As follows, we define an optimization problem and we show the special feature (so-called complementarity principle) of the optimal binary-valued initial conditions.
Vulnerability analysis of climate change impacts in the city of Prague
Lorencová, Eliška ; Emmer, Adam ; Geletič, Jan ; Vačkář, David
Climate change is one of the key challenges of the 21st century, both in terms of adaptation as well as mitigation. The aim of this research was, following the Adaptation Strategy of the City of Prague, to prepare the background analysis for the Adaptation Action Plan, focusing on vulnerability assessment. The vulnerability asssessment focused on the climate change impacts related to: (i) temperature extremes - heatwaves, (ii) insufficient rainwater retention and extreme rainfall. The approach included spatially-specific analysis using ArcGIS based on climatic, land use and socio-economic indicators for the current status and future RCP 4.5 and RCP 8.5 scenarios. Regarding vulnerability to heatwaves, the most affected areas are located in the city center (Prague 2, Prague 3, Prague 6, Prague 7, Prague 1) and some peripheral areas with industrial buildings (e.g. Libeň or Štěrboholy). Vulnerability to extreme precipitation and insufficient rainwater retention was highest particularly at the confluence of the Vltava and Berounka (Velká Chuchle, Prague 16, Zbraslav and Lipence).
Web Disinformation Detection - Case Study - Novicok in Czechia
Řimnáč, Martin
The paper presents a case study of the propaganda usage on a real cause of double agent Sergei Skripal. The formal model describing statements published in web articles is announced and particular interesting aspects of used disinformation are provided together with the reasons, why the disinformation is published. The paper is aimed at the presentation of the data collection to have been created and provides a brief discussion on the used propaganda techniques.
Transforming hierarchical images to program expressions using deep networks
Křen, Tomáš
We present a technique describing how to effectively train a neural network given an image to produce a formal description of the given image. The basic motivation of the proposed technique is an intention to design a new tool for automatic program synthesis capable of transforming sensory data (in our case static image, but generally a phenotype) to a formal code expression (i.e. syntactic tree of a program), such that the code (from evolutionary perspective a genotype) evaluates to a value that is similar to the input data, ideally identical. Our approach is partially based on our technique for generating program expressions in the context of typed functional genetic programming. We present promising results evaluating a simple image description language achieved with a deep network combining convolution encoder of images and recurrent decoder for generating program expressions in the sequential prefix notation and propose possible future applications.
Fulltext: Download fulltextPDF
Detailní simulace proudění, teplot a znečištění vzduchu pro oblast Praha-Dejvice
Resler, Jaroslav ; Geletič, Jan ; Krč, Pavel ; Eben, Kryštof
Simulations of Prague quarter Dejvice were performed with newly developed urban climate model PALM-4U based on LES model PALM. The modelling domain has extent 1000 x 800 m and the resolution of the model was 2 m. Two 24 hours episodes were simulated. The summer episode was intended to assess mainly the UHI effects and the winter episode to assess mainly the air quality issues. Two variants were simulated - the current real situation and the scenario with considered new buildings in the area of Victory Square (Vítězné náměstí). Some comments of the ressults are appended at the end of the report.

Institute of Computer Science : 1,617 records found   beginprevious21 - 30nextend  jump to record:
Interested in being notified about new results for this query?
Subscribe to the RSS feed.