6
M. N. Favorskaya et al.
interval parameters. The effectiveness of the considered approaches is confirmed on
several model problems.
Chapter 20 discusses the spectral form of mathematical description for the representation of iterated Stratonovich stochastic integrals of an arbitrary multiplicity [35,
36]. Some invariant relations for expansion coefficients and iterated Stratonovich
stochastic integrals are obtained which can reduce the computational cost. The spectral characteristics may be defined with respect to an arbitrary complete orthonormal
system for the representation and modeling. For expansion coefficients, the tensor
representation is formulated.
Part V focuses on information technology and artificial intelligence and includes
4 chapters.
Chapter 21 presents the fractal oriented approach [37, 38] for the analysis and
design of distributed algorithms. Its aims are to represent the distributed algorithm as
an “elastic object” that transforms dynamically at runtime. The use of the containercomponent model provides the following advantages: the ability to select automatically a distributed configuration, building a visual model of the elastic computing
organization, and evaluating its effectiveness. Container-component model is integrated with the box-counting fractal analysis method and fractal control based on
dynamic sampling of the workload. The example of fractal analysis and programming
of the distributed gradient ascent algorithm is presented.
Chapter 22 analyses the problem of training qualified professionals in the field of
information technology. It is natural to use software tools to support the educational
process. Given the high entry threshold, the authors propose simple and accessible
software tools that allow one to free up teacher time for effective student training
[39]. The proposed solution does not pretend to the completeness, but it makes it
possible to form control materials, conduct control measures of different levels, take
into account the attendance of classes, the dynamics of the educational process, and
maintain interaction with a group of students.
Chapter 23 introduces a system that collects massive amounts of texts from the
Internet, analyzes them, builds the entity-event ontology, and presents it to the enduser as a knowledge base. It can be also viewed as an automatic text corpus processing
method that allows using of classic statistical and data analysis methods by extracting
domain-specific information from text. As extracted knowledge is highly structured
and easily operated, it can be used by such methods without any further reference to
the source texts.
Chapter 24 demonstrates the possibility of using deep learning methods to create
a system for automatic 3D scanned objects classification. The choice of the appropriate deep architecture is based on a comparative analysis of existing SOTA models
executed on different data sets. To select an object of interest from a large-scale space
scan, the authors consider preprocessing methods: noise data filtering, reference plane
deletion, and removing extraneous objects. An algorithm for the descriptive representation of three-dimensional models based on the modification of existing methods
of ray casting is obtained. The possibility of using this descriptive representation to
solve the problem of searching among 3D models and using search results for 3D
scene auto-completion is demonstrated.
M. N. Favorskaya et al.
interval parameters. The effectiveness of the considered approaches is confirmed on
several model problems.
Chapter 20 discusses the spectral form of mathematical description for the representation of iterated Stratonovich stochastic integrals of an arbitrary multiplicity [35,
36]. Some invariant relations for expansion coefficients and iterated Stratonovich
stochastic integrals are obtained which can reduce the computational cost. The spectral characteristics may be defined with respect to an arbitrary complete orthonormal
system for the representation and modeling. For expansion coefficients, the tensor
representation is formulated.
Part V focuses on information technology and artificial intelligence and includes
4 chapters.
Chapter 21 presents the fractal oriented approach [37, 38] for the analysis and
design of distributed algorithms. Its aims are to represent the distributed algorithm as
an “elastic object” that transforms dynamically at runtime. The use of the containercomponent model provides the following advantages: the ability to select automatically a distributed configuration, building a visual model of the elastic computing
organization, and evaluating its effectiveness. Container-component model is integrated with the box-counting fractal analysis method and fractal control based on
dynamic sampling of the workload. The example of fractal analysis and programming
of the distributed gradient ascent algorithm is presented.
Chapter 22 analyses the problem of training qualified professionals in the field of
information technology. It is natural to use software tools to support the educational
process. Given the high entry threshold, the authors propose simple and accessible
software tools that allow one to free up teacher time for effective student training
[39]. The proposed solution does not pretend to the completeness, but it makes it
possible to form control materials, conduct control measures of different levels, take
into account the attendance of classes, the dynamics of the educational process, and
maintain interaction with a group of students.
Chapter 23 introduces a system that collects massive amounts of texts from the
Internet, analyzes them, builds the entity-event ontology, and presents it to the enduser as a knowledge base. It can be also viewed as an automatic text corpus processing
method that allows using of classic statistical and data analysis methods by extracting
domain-specific information from text. As extracted knowledge is highly structured
and easily operated, it can be used by such methods without any further reference to
the source texts.
Chapter 24 demonstrates the possibility of using deep learning methods to create
a system for automatic 3D scanned objects classification. The choice of the appropriate deep architecture is based on a comparative analysis of existing SOTA models
executed on different data sets. To select an object of interest from a large-scale space
scan, the authors consider preprocessing methods: noise data filtering, reference plane
deletion, and removing extraneous objects. An algorithm for the descriptive representation of three-dimensional models based on the modification of existing methods
of ray casting is obtained. The possibility of using this descriptive representation to
solve the problem of searching among 3D models and using search results for 3D
scene auto-completion is demonstrated.