ARTIFICIAL INTELLIGENT INTRUSION DETECTION SYSTEMS: PERSPECTIVES OF INNOVATIVE TECHNOLOGIES
Оценка 4.7
Занимательные материалы +4
docx
информатика +2
Взрослым
07.11.2019
The most popular development tools of the quantum cryptography technology we compared, the structure and the basic principles of its work is considered. In article, the significance in the modern information society of the information resources requiring safe methods of protection against NSD have discussed. The structure and the basic principles of technology in the field of artificial intellectual system for detection of invasions and also a tendency of development of advanced developments innovative technologies of the system of detection of invasions is considered. Also describes the structure and basic principles of quantum cryptography technology based on properties of quantum systems. Quantum information are a physical quantity characterizing changes occurring in the system during the interaction between the information flow and the external environment. The exploratory development we been experimentally realized by the authors at the facility for multi-channel data transmission network within the grant program. The method of increase in level of information security and protection of confidential information by development innovative were been offered technology of the system of detection of invasions. In the article a method of increasing the level of information security and the protection of confidential information by the quantum artificial intellectual systems of quantum channels communication are considered.
Когнитивные информационные технологии в системах управления.docx
УДК 004.56
DOI 10.25559/SITITO.2017.3.489
Aktayeva A.1,4, Dautov A.1, Niyasova A.3, Gagarina N.4, Bizhigitova D.4, Kussainova U.1, Shatenova G.5
1 Abai Myrzakhmetov Kokshetau University, Kokshetau Kazakhstan
2 SH.Ualikhanov Kokshetau State University, Kokshetau Kazakhstan
3 L. Gumilyiev Eurasian National University, Almaty, Kazakhstan
4 Almaty Technological University, Almaty, Kazakhstan
5 Kazakh Academy of Transport and Communications named after M. Tynyshpaev, Almaty, Kazakhstan
ARTIFICIAL INTELLIGENT INTRUSION DETECTION SYSTEMS: PERSPECTIVES OF
INNOVATIVE TECHNOLOGIES
Abstract
The most popular development tools of the quantum cryptography technology we compared, the structure and
the basic principles of its work is considered. In article, the significance in the modern information society of the
information resources requiring safe methods of protection against NSD have discussed. The structure and the basic
principles of technology in the field of artificial intellectual system for detection of invasions and also a tendency of
development of advanced developments innovative technologies of the system of detection of invasions is considered.
Also describes the structure and basic principles of quantum cryptography technology based on properties of quantum
systems. Quantum information are a physical quantity characterizing changes occurring in the system during the
interaction between the information flow and the external environment. The exploratory development we been
experimentally realized by the authors at the facility for multichannel data transmission network within the grant
program. The method of increase in level of information security and protection of confidential information by
development innovative were been offered technology of the system of detection of invasions. In the article a method of
increasing the level of information security and the protection of confidential information by the quantum artificial
intellectual systems of quantum channels communication are considered.
Keywords
Corporate information systems and network intrusion detection systems; information security; software and
hardware; antivirus systems; artificial intellectual systems.
Introduction
Now the quantum informatics represents the new, quickly developing branch of science connected to use of
quantum technologies for implementation of essentially new methods of info telecommunication and computation:
quantum information, quantum informatics, quantum communication links, quantum cryptography, quantum computer
and artificial neural networks.
Quantum information is the physical quantity characterizing the changes happening in system in case of
interaction of an information flow to an external surrounding.
Quantum information is a new type of information, which can be transfer, but it is impossible to multiply.
Quantum a vector in twodimensional complex vector space also represents twolevel quantum system. Ions, atoms,
electrons, photons, spins of atomic kernels, structures from superconductors and many other physical systems can act as
qubits [8].
Quantum bit or qubit — it's a vector of unit length in 2dimensional complex space, in which some base is
recorded (|0), |1)}, and any complex linear combination 0 and 1 can be written as: a|0) +b|1) [8].
Quantum neural networks (QNN) are one of subspecies of neural networks and represent a combination of
classical neural networks and quantum computing (see fig. 1). Principles of creation and basic elements of QNN
The Artificial Neural Networks (ANN) have some attractive features: parallelism of distributed processing, an
error stability, ability to study and generalize the gained knowledge. The property of generalization was be understood as
ability of ANN to generate the correct outputs for input signals that were not consider in training activity. These
properties do ANN by the system of processing of information, which solves difficult multivariate tasks, excessive to
other techniques [17].
Some system can be called neural if in it it is possible to identify, at least, one neuron. The neural system is
quantum if it is capable to realize quantum computing.
There are two principal reasons of interest in quantum neural networks. One connected to arguments that
quantum processes can play an important role in operation of a brain. Other reason was been connected to rapid growth
of quantum computing which main ideas could be quite postponed for neuro computation that would open for them new
opportunities. Now they have the following advantages:
• elimination of a disastrous forgetting thanks to absence of an interference of images;
• solution of linear and indivisible problems onelayer network;
• absence of connections;
• high data processing rate (1010 bits/s);
• diminutiveness (1011 neurons / ммЗ);
• higher stability, reliability;
• exponential storage capacity;
• the best characteristics in case of smaller number of the hidden neurons;
• fast training [17].
These potential advantages of quantum neural networks motivate their development.
There are different prototypes of quantum neural networks. Some of them are very similar to the classical
analogs while quantum operators who have no classical equivalents, for example, phase shifts, use others. Distinguish a
wide range of different structures of QNN. Different researchers use own analogies for link establishment between a
quantum mechanics and artificial neural networks. The basic concepts of these two areas were give in table 1.
Table 1 – Analogy between a quantum mechanics and artificial neural networks
Classical neural networks
Quantum neural networks
Neuron status
Communication
Learning rule
ix
0,1
p
1
ij
1
s
x x
i
s
j
w
ij
p
s
1
Qubits
x=a 0+ b 1
Complexity
x0 x1 … x p1
Superposition of statuses
of complexity
p
s
1
s
a x
0
s
...
s
x
1
p
Search of the winner
n
max arg(
i
f
)i
Interference as
U
:
' Output Result
N
unitary conversion
Decoherence
(measurement)
s
a x
s
x
k
s
Analogy establishment also is one of the main tasks of the theory of quantum neural networks. The efficiency of
use of neural networks is connected to massive parallel distributed processing of information and nonlinearity of
conversion of vectors of inputs neurons. Besides, quantum systems have much more powerful quantum parallelism, the
expressing superposition principle [17].
QNN works also, as well as classical ANN which consists of several layers of perseptron an input layer, 1 or
several buried layers and an output layer. Each layer is completely connected to the previous layer. Each buried layer
calculates the weighed amount of outputs of the previous layer. If this amount exceeds threshold value, the node
transfers above, otherwise it remains lower. The output layer does the same, as the buried layers, except check of
accuracy. The network in general calculates function by check of the maximum output bit [17].
Determination of the principle of operation of quantum neuron: it obtains input signals or basic data or output
signals of other neurons of QNN through several input channels. Each input signal passes through the connection having
a certain intensity or weight; this weight corresponds to synoptic activity of neuron. A certain threshold value is
connected to each neuron. The weighed amount of inputs is calculated, the threshold value is subtracted from it and as a
result value of activation of neuron turns out and is called the postsynaptic potential of neuron PSP.
The mathematical model of a signal of activation of quantum neuron will be transformed by means of function
of activation and as a result the output signal of neuron turns out (see fig. 3).
Mathematical model of quantum neuron
, where w^j – it is matrix 2x2, act on basis
{|0), |1)}; F^ operator, which can carry out network functioning of quantum cells.
y
F
^
N
1
j
w j xj
^
The present the purpose and the task of the system of detection of invasions are not only monitoring and audit
of events of the information processes proceeding in SIS&Q and also the analysis and prediction of these events in
search of signs of violation of a policy of an information security system. Depending on a source of detection of
invasions distinguish the following subsystems (see fig. 4). The class of tasks which can be solved by means of QNN, is defined by how the network works, and how it
studies. By operation of QNN accepts values of input variables and gives values of output variables. Thus, the network
can be applied in a situation when there is certain known information and it is necessary to obtain from it some yet not
known information. Some examples of such tasks in the field of information security, which can be solved by means of
QNN:
image identification and classification;
decisionmaking and control;
clustering;
prediction;
approximation.
The first level identifies invasions, analyzing events and the traffic arriving on the separate computer while the
second researches a network traffic. The systems of level of applications are located between web and SQLservers. In
a figure 5, the alternative structure of computer system and a network with the System of Detection of Invasions (SDI) of
the organization was been provided [24].
In turn, as practice, systems of detection of invasions, depending on the used technology of detection of
different types of the attacks and threats shows, divide into two main classes (see fig. 5) [2].
The firs guided by model of malicious behavior (template/signature) and compare model to an event stream of
information process. Based on the comparative analysis the system makes the decision to lock this or that packet, or to
pass for interaction directly with infrastructure SIS&Q.
Signature
of invasions have
and efficiency of
rather
low
hardware. The read
signature systems
invasions:
•
automatic input of
• absence
of
blocking
signatures;
systems of detection
high performance
detection in case of
requirements to the
by shortcomings of
of detection of
impossibility of
new signatures;
of
unknown
system of
• absence of opportunities of prediction of actions of the malefactor;
• absence of a subsystem for monitoring and audit of hardware resources, etc.
Principal a lack of the existing signature systems of detection of invasions that they cannot profile the
intercepted data stream of the distributed invasions for their classification and framing of a signal of invasion [1]. Different types of heuristic algorithms are usually applied to recognition and detection of invasions: combined
systems with heuristic scanning. This method based on signatures and heuristics is designed to improve ability of SDI to
apply signatures and to recognize the modified versions of invasions. However, this technology, is applied very carefully
as can increase the number of false operations [1].
Systems of detection of invasions of the second class are designed on the basis of models of normal behavior of
a template and look for the abnormal entrances in an event stream of information process for recognition of unknown
invasions. The research problem of normal and aberrant behavior of info telecommunication systems is very difficult
and complex. Therefore, erudite practicians mark that the following parameters can be select by criteria of the analysis
and assessment of the SDI methods:
1) level of observation over system;
2) method verifitsiruyemost (expert assessment of correctness of a method in use systems of detection of
invasions);
3) adaptivity of a method (resistance of a method to small changes in case of attack implementation);
4) accuracy of detection of invasions and level of false operations and etc. [24].
Problem definition of a research
According to materials of the seventh International conference on systematization of computer viruses of
"VIRUS 97" in 1997, the artificial immune system (AIS) for a cyberspace within the project of IBM were been realized.
In a figure 7, the diagram of the artificial immune computing system was provided [24].
This is model functioning of AIS is based on basic provisions and mechanisms of biological immune system
(generation and detectors, selection of unwanted detectors, cloning and a mutation of detectors, formation of immune
memory). The generalized diagram of the mechanism of operation of AIS, given in a figure 8. The main opportunities of
AIS of the offered model:
1) the continuous updating of immune detectors;
2) training of immune detectors is correct to classify unknown images;
3) selfsufficiency of immune detectors;
4) the mechanism of "the programmed death";
5) cloning and mutation;
6) existence of immune memory of the adaptive system of detection of invasions.
In implemented the project of IBM on creation of artificial immune system the following stages are
included:
1) detection of unknown viruses: congenital immunity;
2) data collection of a sample of viruses (separation and transfer)
3) vaccine framing (updated antivirus basis);
4) the adaptive immune system;
5) delivery updating and distribution of a basis of vaccination (antibodies) [24]. The Artificial Immune System (AIS) shall contain components of both congenital and adaptive immune
protection. By analogy with congenital immunity, it shall have the generalized mechanisms of recognition of
harmful changes, but it is not enough.
accidental
generation
of detectors
training and
selection of
detectors
destruction of
"bad" detectors
functioning of detectors in system
destruction of
detectors
detection of
anomaly
formation of immune
memory (framing of
antibodies)
cloning, mutation of
detectors
Fig. 8. The generalized diagram of the mechanism of operation of the
AIS project IBM antivirus [24]
As well as the adaptive immunity of the biological AIS system shall have specific mechanisms of recognition
and detection of invasions in the arsenal. Integrating these very general provisions with a row of reasons of practical
character, it was formulated a necessary set of requirements that shall be executed for effective preventing of invasions:
1) congenital immunity;
2) the adaptive immunity;
3) highspeed performance;
4) modular building; 5) safety and reliability;
6) safety;
7) user monitoring. [24]
Premise for creation of effective systems of detection of invasions is development of the artificial immune
systems (AIS) and neural network technologies (INS) which have biological bases. But the majority of the invasions
which are available today artificial immune detection systems meets only a small part of the listed requirements.
However development and deployment of the stable endtoend system of information security (ETESIS) from quickly
spreaded invasions represents the difficult task and requires observance of all listed requirements, any of nowadays
existing systems does not provide such protection level [24].
Results of researches
The detail analysis of foreign development of AIS and INS allowed to select the main perspective directions of
development of effective systems of detection of invasions: neuropackets, neural network expert systems, antivirus
programs with switching on of AIS and INS of algorithms. The carriedout analysis of literary and open sources shows
absence of the finished decisions in this direction. Therefore the relevant task is development of effective algorithms and
methods of hybrid systems of detection of invasions which allows to find unknowns and is based on integration of
artificial immune systems and neural network technologies on the basis of evolutionary programming. We read that the
ability of such systems to training and generalization of results allows creating hybrid intelligent systems of detection of
unknown invasions.
Due to the intensive development of innovative technologies researches in electronics, creation of intellectual
software and hardware products of applicationoriented informatics and quantum technologies are of particular
importance. The quantum information and technologies based on its unusual properties in the future will affect bases and
further development of information space, and broad use of quantum technologies assumes a scientific and technological
revolution, which scales it, is very difficult to provide. Distribution of technology of quantum communication is one of
perspective and at the same time real steps in strategic plans of a row of the countries of Europe and the USA, Japan.
The theory of quantum information will cardinally change the modern views of scientific community to a basis
of an information security system. Carrying out experiments and researches on support of information security is of great
scientific interest on search of the solution of the main objectives and problems facing quantum cryptography systems:
the task of detection of single photons with high probability in the given quantum status in case of the low level of false
operations, absence of controlled sources of single photons, a problem of increase in range of transmission and small
generation rate of a quantum key.
Use of quantum technologies in the field of support of an information security system one of the most
paradoxical manifestations of quantum technology attracting in recent years huge interest of experts. First, by
transmission of the ciphered messages on two more communication links quantum and traditional. Quantum neuronets
is one of the most promptly developing applicationoriented directions of quantum informatics, and provides informing
on attempt of interception of the transmitted data.
Inference
Researches in the field of quantum information can lead not only to the positive consequences, but also to the
negative. The quantum cryptography based on application of a quantum neuronet in the future will replace the used
cryptography systems, and will be applied on an equal basis with normal means of an infotelekommunication. The
relevance and scale of the problems connected to support of information security every day will increase, and
development of quantum information will bring the results and, perhaps in the near future, will lead to essential change
of a scientific pattern of the world in the field of IT.
Today the transition tendency from program to hardwaresoftware implementation of neural network algorithms
with sharp increase in number of development of GSI of neurochips with neural network architecture is watched.
According to the open press, researches of Microsoft on creation artificial immune and neural technologies for
monitoring and audit of detection of invasions for future generations of OS are financed. Most likely, it means that the
scope gidridny artificial immune and neural is much wider than technology as the majority of development nevertheless
are secret. Therefore we read that it is necessary to conduct very intensive and largescale researches of fundamental and
applicationoriented character for the decision of the task of support of a reliable endtoend system of information
security in information security field. Broad application of quantum information technologies assumes a scientific and
technological revolution which scales now it is even difficult to provide [8]. 344 p.
1. Artificial immune systems and their application / Under the editorship of D. Dasgupta. M.: Fizmatlit, 2006.
2. Bezobrazov, S.V., Golovko V. A. Artificial immune systems for protection of information: detection and
classification of computer viruses//Materials of Vseross. Naych.confernce Neyroinformatika, MEPhI, Moscow,
2023 January. 2008. Moscow, 2008. 2328
pp.
3. Bezobrazov S.V. Artificial immune systems for information security: application of LVQ
network//Neuroinformatics2007: materials IX of Vseross. science tech. conference, Moscow, MEPhI, 2007,
2735 pp.
4. Bezobrazov S.V., Golovko V. A. Artificial immune systems for information security: detection and
classification of computer viruses//Neuroinformatics2008. Moscow, MEPhI, 2008. 2327 pp.
5. Aktayeva A. and etc. Security of information: using of quantum technologies//International Journal of Open
Information Technologies. vol 4, №. 4, 2016, 4048 pp. www.injoit.org
6. Aktayeva A.U., Ilipbayeva L.I. Innovative technologies in an information security system: quantum
technologies//Modern innovative technologies and IT education. 2014, vol 1, №1(9), 320326 pp.
7. Sandpiper of Page. Classical cryptography//Photonics 2010,2, 3641 pp.
8. Holevo A. S Mathematical fundamentals of quantum informatics. M.:2016, 125 p.
9. Kurochkin V.L. Experimental installation for quantum cryptography with the single polarized
photons//Magazine of technical physics, 2005, V.75, №6
10. Glushchenko L.A., Morgunov K.K. Cost of information security in laser communication lines// URL:
http://www.oop ros.org/maket2012/part1/ref1_2/1.3.2.pdf (circulation date 01.10.2017)
11. D. F. lugger. Artificial intelligence, strategy and methods of the solution of complex problems. M.: Williams,
12. Russell S. and etc. Artificial intelligence: modern approach. M.: Williams, 2007, 1408 p.
13. Gorodetski V.I., Kotenko I.V., Karsaev O. Multiagent technologies for computer network security: Attack
simulation, intrusion detection and intrusion detection learning//International Journal of Computer Systems
Science & Engineering. 2003. № 4,191200 pp.
14. www.comsec.spb.ru
15. http://www.securitylab.ru/contest/299868.php_ (circulation date 01.10.2017)
16. Aeppli G., Rosenbaum T. Quantum Annealing and Related Optimization Methods. Heidelberg: Springer
2003, 864 p.
Verlag, 2007, vol 679, 159169 pp.
2009, vol 10, № 5, 29612970 pp.
1135711361 pp.
17. Altaisky M., Rao V. Inverted Mexican Hat Potential in Activation of Receptor Cells // Nonlin. Analysis B,
18. Beck F. Synaptic Quantum Tunnelling in Brain Activity//Neuroquantology. 2008, vol 6, №. 2, 140151pp.
19. Beck F., Eccles J. Quantum Aspects of Brain Activity and the Role of Consciousness// PNAS.1992, vol 89,
20. Behera L., Kar I., Elitzur A. A Recurent Quantum Neural Network Model to Describe Eye Tracking of Moving
Targets// Found. Phys. Lett. 2005, vol 18, № 4, 357370 pp.
Поступила: 22.10.2017
Note on the authors:
Aktayeva Alkena, doctor Ph.D, associate professor "Information technologies", Almaty technological university,
ulzhan97@mail.ru
shatenova94@mail.ru
aaktaewa@list.ru
Dautov Aibek, MSc., Ph.D student, dozent, department of Information systems and Informatics, SH. Ualikhanov
Kokshetau State University, d.abeke@mail.ru
Niyazova Rozamgul, Candidate of Technical Sciences, associate professor "A theoretical informatika and information
security", L. Gumilyiev Eurasian National University, rosamgul@list.ru
Gagarina Nadezhda, doctor of PhD, Candidate of Economic Sciences, associate professor "Information technologies",
Almaty technological university, ngagarina@mail.ru
Bizhigitova Danakyz, doctor of PhD, master of IS technology, teacher of Information Technologies department, Almaty
technological university, bdana@mail.ru
Kusainova Ulzhan, master of IS technology, teacher of "ISI", Abai Myrzakhmetov Kokshetau University,
Shatenova Gulmira, undergraduate, Kazakh Academy of Transport and Communications named after M. Tynyshpaev,
ARTIFICIAL INTELLIGENT INTRUSION DETECTION SYSTEMS: PERSPECTIVES OF INNOVATIVE TECHNOLOGIES
ARTIFICIAL INTELLIGENT INTRUSION DETECTION SYSTEMS: PERSPECTIVES OF INNOVATIVE TECHNOLOGIES
ARTIFICIAL INTELLIGENT INTRUSION DETECTION SYSTEMS: PERSPECTIVES OF INNOVATIVE TECHNOLOGIES
ARTIFICIAL INTELLIGENT INTRUSION DETECTION SYSTEMS: PERSPECTIVES OF INNOVATIVE TECHNOLOGIES
ARTIFICIAL INTELLIGENT INTRUSION DETECTION SYSTEMS: PERSPECTIVES OF INNOVATIVE TECHNOLOGIES
ARTIFICIAL INTELLIGENT INTRUSION DETECTION SYSTEMS: PERSPECTIVES OF INNOVATIVE TECHNOLOGIES
ARTIFICIAL INTELLIGENT INTRUSION DETECTION SYSTEMS: PERSPECTIVES OF INNOVATIVE TECHNOLOGIES
ARTIFICIAL INTELLIGENT INTRUSION DETECTION SYSTEMS: PERSPECTIVES OF INNOVATIVE TECHNOLOGIES
Материалы на данной страницы взяты из открытых истончиков либо размещены пользователем в соответствии с договором-офертой сайта. Вы можете сообщить о нарушении.