ARTIFICIAL INTELLIGENCE IN CYBER SECURITY

INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS
www.ijrcar.com
Vol.4 Issue 5, Pg.: 1-5
May 2016
D r . P r a n a v P a t i l Page 1
INTERNATIONAL JOURNAL OF
RESEARCH IN COMPUTER
APPLICATIONS AND ROBOTICS
ISSN 2320-7345
ARTIFICIAL INTELLIGENCE IN
CYBER SECURITY
Dr. Pranav Patil
Assistant Professor, Department of Computer Science, M. J. College, Jalgaon, Maharashtra, India
Abstract: - The speed of processes and also the quantity of knowledge to be utilized in defensive the cyber
area cannot be handled by humans while not sizeable automation. However, it is troublesome to develop software
system with standard mounted algorithms (hard-wired logic on deciding level) for effectively defensive against the
dynamically evolving attacks in networks. This example may be handled by applying strategies of computing that
offer flexibility and learning capability to software system. This paper presents a quick survey of computing
applications in cyber security, and analyzes the prospects of enhancing the cyber security capabilities by suggests
that of accelerating the intelligence of the security systems. Once measuring the papers obtainable regarding AI
applications in cyber security, we will conclude that helpful applications exist already. They belong; initial of all, to
applications of artificial neural nets in perimeter security and a few alternative cyber security areas. From the
opposite facet – it has become obvious that several cyber security issues may be resolved with success only
strategies of AI are getting used. For instance, wide information usage is critical in deciding, and intelligent call
support is one in all however unresolved issues in cyber security.
Keywords: Cyber Security Methods, Artificial Intelligence, Visual nets, Expert Systems.
1. Introduction
It is understandable that security against intelligent cyber bats will be achieved only by intelligent code, and events
of the most recent years have shown quickly increasing intelligence of malware and cyber-weapons. Application of
network central warfare makes cyber incidents particularly dangerous, and changes in cyber security are desperately
needed. The new security ways like dynamic setup of secured perimeters, comprehensive scenario awareness,
extremely machine-driven reaction on attacks in networks would require wide usage of AI ways and knowledge-
based tools. Why has the role of intelligent code in cyber operations accrued thus rapidly? Wanting nearer at the
cyber house, one will see the subsequent answer. AI is required, initial of all, for speedy reaction to things in web.
One should be able to handle great deal of data in no time so as to explain and analyze events that happen in cyber
house and to form needed choices. The speed of processes and also the quantity of information to be used cannot be
handled by humans while not substantial automation. However, it is troublesome to develop code with standard
mounted algorithms (hard-wired logic on deciding level) for effectively defensive against the attacks in cyber house,

www.ijrcar.com
May 2016
as a result of new threats seem perpetually.
2. Concerning AI
Artificial intelligence (AI) as a field of research project (also known as machine intelligence within the beginning) is
sort of as previous as electronic computers are a prospect of building devices/software/systems additional intelligent
than persons has been from the first days of AI “on the horizon”. The matter is that the time horizon moves away
once time passes. We have witnessed the determination of variety of showing intelligence exhausting issues by
computers like enjoying sensible chess, as an example. Throughout the first days of computing the chess enjoying
was thought of a benchmark showing a true intelligence. Even in seventies of the last century, once the pc chess was
on the master’s level, it appeared nearly not possible to form a program that might beat the planet champion. it's
typically accepted that AI will be thought of in 2 ways: as a science aimed toward making an attempt to get the
essence of intelligence and developing typically intelligent machines, or as a science providing ways for
determination complicated issues that can't be solved while not applying some intelligence like, as an example,
enjoying sensible chess or creating right choices supported giant amounts of knowledge. Within the gift paper we
are going to take the second approach, advocate for applying specific AI ways to cyber security issues.
A large range of ways is developed within the AI field for finding laborious issues that need intelligence from
the human perspective. Some of these ways have reached a stage of maturity wherever precise algorithms exist that
is supported these ways. Some ways have even become thus wide known that they are not thought of happiness to
AI any further, but became a section of some application area, as an example, data processing algorithms that have
emerged from the training subfield of AI. It might be impossible to do to offer a lot of or less complete survey of all
much helpful AI methods in a very transient survey. Instead, we have sorted the ways and architectures in many
categories: neural nets, knowledgeable systems, intelligent agents, search, machine learning, data processing and
constraint finding. We define these classes here, and that we provide references to the usage of individual ways in
cyber security. We are not aiming to discuss tongue understanding, artificial intelligence and pc vision that we
contemplate specific applications of AI. Robots and pc vision have positively spectacular military applications,
however we have not found something specific to cyber security there.
2.1. Visual nets: visual nets have an extended history that begins with the invention of perceptron by Frank
Rosenblatt in 1958 – a man-made nerve cell that has remained one among the foremost well-liked components of
neural nets. Already a little variety of perceptrons combined along will learn and solve fascinating issues. However
neural nets will include an oversized variety of artificial neurons. So neural nets offer a practicality of massively
parallel learning and decision-making. Their most distinguished feature is that the speed of operation. They’re well
matched for learning pattern recognition, for classification, for choice of responses to attacks etc. they will be
enforced either in hardware before in software system. Neural nets are well relevant in intrusion detection and
intrusion bar. There are proposals to use them in DoS detection, pc worm detection, spam detection, zombie
detection, and malware classification and in rhetorical investigations. A reason for the recognition of neural nets in
cyber security is their high speed, if enforced in hardware or utilized in graphic processors. There are new
developments within the neural nets technology: third generation neural nets prickling neural networks that imitate
organic neurons a lot of realistically, and supply a lot of application opportunities.
2.2Expert systems: These are unquestionably the foremost wide used AI tools. Associate skilled system is software
system for locating answers to queries in some application domain bestowed either by a user or by another software
system. It will be directly used for 98 call support, e.g. in diagnosing, in finances or in computer network. There’s a
good sort of skilled systems from little technical diagnostic systems to terribly massive and hybrid systems for
finding complex issues. Conceptually, associate skilled system includes a mental object, wherever skilled
information a few specific application domains are hold on. Besides the mental object, it includes associate illation
engine for account answers supported this information and, possibly, further information a few state of affairs.
Empty mental object and illation engine are along referred to as skilled system shell - it should be stuffed with
information, before it will be used. This system shell should be supported by software system for adding information

www.ijrcar.com
May 2016
within the mental object, and it will be extended with programs for user interactions, and with different programs
that will be utilized in hybrid skilled systems. Developing associate skilled system means that, first,
selection/adaptation of associate skilled system shell and, second, exploits skilled information and filling the mental
object with the information. The second step is out and away a lot of difficult and time overwhelming than the
primary. There are several tools for developing expert systems. In general, a tool includes associate expert system
shell and has conjointly a practicality for adding information to the information repository. Expert systems will have
further practicality for simulation, for creating calculations etc. There are many various information illustration
forms in expert systems; the foremost common may be a rule-based illustration. However the utility of associate
expert system depends principally on the standard of information within the expert system’s knowledge domain, and
not most on the inner type of the information illustration. This leads one to the information acquisition drawback
that is crucial in developing real applications. Example of a Cyber Security expert system is one for security
designing. This expert system facilitates significantly choice of security measures, and provides guidance for best
usage of restricted resources. There are early works on mistreatment expert systems in intrusion detection.
2.3. Intelligent agents: Intelligent agents are software system elements that possess some options of intelligent
behavior that produces them special: pro-activeness, understanding of an agent communication language, reactivity
(ability to form some selections and to act). They will have a designing ability, quality and reflection ability. Within
the software system engineering community, there is a thought of software system agents wherever they are thought
of to be objects that are a minimum of proactive and have the flexibility to use the agent communication language.
comparison agents and objects, one will say that objects is also passive, and that they do not need to perceive any
language victimization intelligent agents in security against DDoS has been represented, wherever simulation shows
that cooperating agents will effectively defend against DDoS attacks. Once determination some legal and conjointly
industrial several issues, it should be attainable in premise to develop a “cyber police” subsisting of mobile
intelligent agents. This may need implementation of infrastructure for supporting the cyber agents’ quality and
communication, however should be inaccessible for adversaries. This may need cooperation with ISP-s. Multi-agent
tools will give a lot of complete operational image of the cyber house, as an example, a hybrid multi-agent and
neural network-based intrusion detection method has been projected. Agent-based distributed intrusion detection is
represented.
2.4. Search: Search may be a universal technique of downside finding which will be applied altogether cases once
no different ways of downside finding are applicable. Individuals apply search in their daily life perpetually, while
not listening to it. Little should be known so as to use some general search formula within the formal setting of the
search problem: one should be able to generate candidates of solutions, and a procedure should be out there for
deciding whether or not a planned candidate satisfies the wants for an answer. However, if extra information may be
exploited to guide the search, then the potency of search may be drastically improved. Search is gift in some type
nearly in each intelligent program, and its potency is commonly vital to the performance of the total program. An
excellent form of search ways are developed that take under consideration the precise information regarding
particular search issues. Though several search ways are developed in AI, and that they are wide utilized in several
programs, it's rarely thought-about because the usage of AI. For instance, dynamic programming is actually utilized
in finding optimum security issues, the search is hidden within the package and it's not visible as an AI application.
Search on and or trees, αβ-search, minimax search and random search square measure wide utilized in games
package, and that they are helpful in decision-making for cyber security. The αβ-search formula, originally
developed for pc chess, is an implementation of a typically helpful preparation of “divide and conquer” in problem
finding, and mostly in deciding once 2 adversaries are selecting their absolute best actions. It uses the estimates of
minimally secured win and maximally doable loss. This allows one typically to ignore great amount of choices and
significantly to hurry up the search.
2.5. Learning: Learning is raising a data system by extending or rearranging its cognitive content or by raising the
illation engine. This is often one in all the foremost fascinating issues of AI that is beneath intensive investigation.

www.ijrcar.com
May 2016
Machine learning includes procedure strategies for getting new data, new skills and new ways that to prepare
existing data. Issues of learning vary greatly by their complexness from easy constant learning which suggests
learning values of some parameters, to difficult kinds of symbolic learning, for instance, learning of ideas,
grammars, functions, even learning of behavior. AI provides strategies for each -- supervised learning further as
unattended learning. The latter is very helpful within the case of presence of enormous quantity of knowledge, and
this is often common in cyber security wherever giant logs will be collected. Data processing has originally adult out
of unattended learning in AI. Unattended learning will be a practicality of neural nets, especially, of self-organizing
maps. A distinguished category of learning strategies is implanted by parallel learning algorithms that are
appropriate for execution on parallel hardware. These learning strategies are diagrammatical by genetic algorithms
and neural nets. Genetic algorithms and symbolic logic has been, as an example, utilized in threat detection systems
represented.
2.6.Constraint finding: Constraint finding or constraint satisfaction may be a technique developed in AI for
locating solutions for issues that area unit conferred by giving a group of constraints on the answer, e.g. logical
statements, tables, equations, inequalities. An answer of a drag may be a assortment of values that satisfy all
constraints. Actually, there are many various constraint determination techniques, betting on the character of
constraints. On a really abstract level, nearly any downside will be conferred as a constraint satisfaction downside.
Particularly, several designing issues will be conferred as constraint satisfaction issues. These issues are troublesome
to resolve as a result of great amount of search required normally. All constraint determination strategies are aimed
toward limiting the search by taking into consideration specific info regarding the actual category of issues.
Constraint determination will be used in scenario analysis and call support together with logic programming.
3. Challenges in Intelligent Cyber Security
When coming up with the long run analysis, development and application of AI ways in Cyber Security, one needs
to distinguish between the immediate goals and long views. There are varied AI ways directly applicable in Cyber
Security, and present are immediate Cyber Security issues that must a lot of intelligent solutions than are enforced
nowadays. As yet we have mentioned these existing immediate applications. Within the future, one will see
promising views of the appliance of fully new principles of data handling in state of affairs management and
deciding. These principles embrace introduction of a standard and hierarchal data design within the deciding
software system. This sort of design has been planned. A difficult application space is that the data management for
internet central warfare. Only automatic data management will guarantee fast state of affairs assessment that
provides a choice superiority to leaders and decision manufacturers on any C2 level. Knowledgeable systems are
already getting used in several applications, typically hidden within an application, like within the security measures
coming up with software system. However, knowledgeable systems will get wider application, if massive data bases
are going to be developed. This may need tidy investment in data acquisition, and development of huge standard
data bases. Considering a lot of distant future -- a minimum of some decades ahead, maybe we should always not
prohibit us to the “narrow AI”. Some individuals are convinced that the grand goal of the AI development of
artificial general intelligence will be reached within the middle of the current century. The primary conference on
artificial general intelligence was control in 2008 at the University of Memphis. The Singularity Institute for AI,
supported in 4000, warns researchers of a danger that exponentially quicker development of intelligence in
computers could occur. This development could result in Singularity, delineate in follows: “The Singularity is that
the technological creation of smarter-than-human intelligence. There are many technologies that are usually
mentioned as heading during this direction. The foremost usually mentioned is perhaps AI; however there are others
many totally different technologies that, if they reached an intensity of sophistication, would change the creation of
smarter-than-human intelligence. An opportunity that includes smarter-than-human minds is actually totally different
in a very manner that goes on the far side the standard visions of a future stuffed with advanced devices.” A
researcher has expected the event to come back up with Singularity. One needn't to believe the Singularity threat;
however the fast development of knowledge technology will certainly change one to make significantly higher

www.ijrcar.com
May 2016
intelligence into software system in coming back years. Severally of whether or not the factitious general
intelligence is obtainable or Singularity comes, it's crucial to own the power to use higher AI in cyber security than
the offenders have it.
4. Conclusions
In the present scenario of quickly growing intelligence of malware and class of cyber-attacks, it is inescapable to
develop intelligent cyber security ways. The expertise in DDoS mitigation has shown that even a security against
large-scale attacks will be undefeated with rather restricted resources once intelligent ways are used. An analysis of
publications shows that the AI results most generally applicable in cyber security are provided by the analysis in
artificial visual nets. Applications of visual nets can keep on in cyber security. There is additionally an imperative
would like for application of intelligent cyber security ways in many areas wherever neural nets are not the foremost
appropriate technology. These areas are called support, scenario awareness and data management. Professional
system technology is that the most promising during this case. It is not clear however fast development of general
computing is ahead, however a threat exists that a replacement level of computing could also be utilized by the
attackers, as presently because it becomes obtainable. Obviously, the new developments in knowledge
understanding, illustration and handling furthermore in machine learning can greatly enhance the cyber security
capability of systems that will use them.
REFERENCES:
[1] E. Tyugu. Algorithms and Architectures of Artificial Intelligence. IOS Press. 2007.
[2] B. Mayoh, E. Tyugu, J. Penjam. Constraint Programming. NATO ASI Series, v. 131, Springer Verlag. 1994.
[3] I. Bratko. PROLOG Programming for Artificial Intelligence. Addison-Wesley, 2001 (third edition).
[4] http://singinst.org/overview/whatisthesingularity/
[5] F. Rosenblatt. The Perceptron -- a perceiving and recognizing automaton. Report 85-460-1, Cornell
Aeronautical Laboratory, 1957.
[6] U. Schade, M. R. Hieb. A Battle Management Language for Orders, Requests and Reports. In: 2007 Spring
Simulatin Ineroperability Workshop. Norfolk, USA, 2006
[7] F. Barika, K. Hadjar, and N. El-Kadhi, “Artificial neural network for mobile IDS solution,” in Security and
Management, 2009.
[8] P. Norvig, S. Russell. Artificial Intelligence: Modern Approach. Prentice Hall, 2000.
[9] http://en.wikipedia.org/wiki/Expert_system. Expert System. Wikipedia.
[10] http://en.wikipedia.org/wiki/Conficker
[11] TF. Lunt, R. Jagannathan. A Prototype Real-Time Intrusion-Detection Expert System.Proc. IEEE
Symposium on Security and Privacy, 1988, p. 59.
[12] V. Chatzigiannakis, G. Androulidakis, B. Maglaris. A Distributed Intrusion Detection Prototype Using
Security Agents. HP OpenView University Association, 2004.
[13] R. Kurtzwell. The Singularity is Near. Viking Adult. 2005.
[14] J. Kivimaa, A. Ojamaa, E. Tyugu. Graded Security Expert System. Lecture Notes in Computer Science, v.
5508. Springer, 2009, 279-286.
[15] J. Kivimaa, A. Ojamaa, E. Tyugu. Pareto-Optimal Situation Analysis for Selection of Security Measures.
Proc. MilCom, 2008.
[16] B. Iftikhar, A. S. Alghamdi, “Application of artificial neural network in detection of dos attacks,” in SIN
’09: Proceedings of the 2nd international conference on Security of information and networks. New York,
NY, USA: ACM, 2009, pp. 229–234.
[17] P. Salvador et al. Framework for Zombie Detection Using Neural Networks. In: Fourth International
Conference on Internet Monitoring and Protection ICIMP-09, 2009.
[18] J. Bai, Y. Wu, G. Wang, S. X. Yang, and W. Qiu, A novel intrusion detection model based on multi-layer
self-organizing maps and principal component analysis, in Advances in Neural Networks. Lecture Notes in
Computer Science. Springer, 2006.

ARTIFICIAL INTELLIGENCE IN CYBER SECURITY

Recommended

Recommended

More Related Content

Similar to ARTIFICIAL INTELLIGENCE IN CYBER SECURITY

Similar to ARTIFICIAL INTELLIGENCE IN CYBER SECURITY (20)

More from Cynthia King

More from Cynthia King (20)

Recently uploaded

Recently uploaded (20)

ARTIFICIAL INTELLIGENCE IN CYBER SECURITY