|
| |
|
| |
|
|
Using Artificial Neural Networks to Study Non-Linear Properties of Single Neurons in the Auditory Cortex
Hromádka, Tomáš ; Kůrková, Věra (vedoucí práce) ; Maršálek, Petr (oponent)
Neurony v naších mozcích převádějí informace o okolním světě do elektrických akčních potenciálů. Klíčem k pochopení funkce neuronů, a poté neuronových sítí a mozku, je vědět, které části okolního světa jsou zastoupeny v aktivitě neuronů, a jakým způsobem je tato informace kódována. Tradiční metodou zkoumání funkce neuronu je zaznamenání odpovědi neuronu na podněty různé složitosti a následně snaha vysvětlit přenosovou funkci neuronu pomocí lineárního modelu. Většina neuronů však obsahuje nelineární přenosovou funkci. V této práci jsme použili umělé neuronové sítě ke studiu nelineárních přenosových funkcí neuronů sluchové kůry. Zaznamenali jsme podprahovou aktivitu neuronů jako odpověď na stimulaci přirozenými zvuky a poté jsme použili neuronové sítě v roli nelineárních aproximátorů k odhadu přenosových funkcí neuronů. Umělé neuronové sítě úspěšně aproximovali přenosové funkce a v průměru odhadli funkce neuronů nejméně stejně dobře jako lineární modely. Powered by TCPDF (www.tcpdf.org)
|
|
|
ITAT 2014. Information Technologies - Applications and Theory. Part II
Kůrková, Věra ; Bajer, Lukáš ; Peška, L. ; Vojtáš, P. ; Holeňa, Martin ; Nehéz, M.
ITAT 2014. Information Technologies - Applications and Theory. Part II. Prague : Institute of Computer Science AS CR, 2014. 145 p. ISBN 978-80-87136-19-5. This volume is the second part of the two-volume proceedings of the 14th conference Information Technologies – Applications and Theory (ITAT 2014), which was held in Jasná, Demänovská Dolina, Slovakia, on September 25–29, 2014. ITAT is a computer science conference with the primary goal of exchanging information on recent research results. Overall, 51 papers were submitted to all conference tracks. This volume presents papers from the workshops and an extended abstract of a poster. Three specialized workshops were held as a part of the conference: Data Mining and Preference Learning on Web, Computational Intelligence and Data Mining, and Algorithmic Aspects of Complex Networks Analysis.
|
|
|
ITAT 2014. Information Technologies - Applications and Theory. Part I
Kůrková, Věra ; Bajer, Lukáš
ITAT 2014. Information Technologies - Applications and Theory. Part I. Prague : Institute of Computer Science AS CR, 2014. 101 p. ISBN 978-80-87136-18-8. This volume is the first part of the two-volume proceedings of the 14th conference Information Technologies – Applications and Theory (ITAT 2014). The conference was held in Jasná, Demänovská Dolina, Slovakia, on September 25–29, 2014. ITAT is a computer science conference with the primary goal of exchanging information on recent research results between Czech and Slovak scientific communities, and it presents a platform for young researchers and PhD students to start new collaborations. This year, it was held in parallel with two collocated conferences Datakon and Znalosti with which it shared some invited plenary talks and a poster session. Overall, 51 papers were submitted to all conference tracks. This volume presents 16 papers of the main track, which were selected by the program committee based on at least two reviews by the program committee members. Papers from the three workshops and extended abstracts of posters are included in the second volume.
|
|
|
Representations of Boolean Functions by Perceptron Networks
Kůrková, Věra
Limitations of capabilities of shallow perceptron networks are investigated. Lower bounds are derived for growth of numbers of units and sizes of output weights in networks representing Boolean functions of d variables. It is shown that for large d, almost any randomly chosen Boolean function cannot be tractably represented by shallow perceptron networks, i.e., each its representation requires a network with number of units or sizes of output weights depending on d exponentially
|
|
|
Capabilities of Radial and Kernel Networks
Kůrková, Věra
Originally, artificial neural networks were built from biologically inspired units called perceptrons. Later, other types of units became popular in neurocomputing due to their good mathematical properties. Among them, radial-basis-function (RBF) units and kernel units became most popular. The talk will discuss advantages and limitations of networks with these two types of computational units. Higher flexibility in choice of free parameters in RBF will be compared with benefits of geometrical properties of kernel models allowing applications of maximal margin classification algorithms, modelling of generalization in learning from data in terms of regularization, and characterization of optimal solutions of learning tasks. Critical influence of input dimension on behavior of these two types of networks will be described. General results will be illustrated by the paradigmatic examples of Gaussian kernel and radial networks.
|
|
| |
|
| |
|
|
Gaussian Radial and Kernel Networks with Varying and Fixed Widths
Kůrková, Věra
The role of widths of Gaussians in computational models which they generate is investigated. Suitability of Gaussian kernel models with fixed widths for regression is proven in terms of their universal approximation capability. Large sets of argminima of error functionals minimized during learning from data over Gaussian networks with varying widths are described. Dependence of stabilizers modelling generalization on widths of Gaussian kernels and the input dimension is estimated.
|
host ::
RSS.