International Journal of Advance Research and Innovation


Application of Neural Networks to Intrusion Detection



Yüklə 0,71 Mb.
Pdf görüntüsü
səhifə8/13
tarix27.01.2023
ölçüsü0,71 Mb.
#81145
1   ...   5   6   7   8   9   10   11   12   13
IJARI-ME-14-09-106 (1)

Application of Neural Networks to Intrusion Detection:  
Intrusion Detection System are becoming largely employed 
as a fundamental security systems to secure company 
networks. Ideally, an IDS has the capacity to detect in real-
time all (attempted) intrusions, and to execute work to stop 
the attack (for example, modifying firewall rules). 
Commercial tools available today have limitations in 
detecting real intrusions, and neural network is an efficient 
way to improve the performances of IDS systems which are 
based on the misuse detection model and the anomaly 
detection model. 
Neural Network in Communication: Typical RF and 
wireless circuits comprise a large number of linear and non-
linear components. Complexity of RF portion of a wireless 
system continues to increase in order to support multiple 
standards, multiple frequency bands, need for higher 
bandwidth and stringent adjacent channel specifications. 
The time required to carry out a virtual prototyping of such 
complex circuits and their trade-off analysis with the 
baseband circuitry, can be unacceptably long, because both 
the circuit simulation and optimization procedures can be 
very time-consuming. Typically, one divides the task into 
that of designing the non-linear elements or sub-circuits at 
module level. Neural network speeds up the modeling RF 
circuits. 
Neural Network in Robotics: A novel topological world 
model and region-filling algorithm is used for autonomous 
vacuuming robots. Humans, and other animals, model their 
environment using topologies of landmarks. This type of 
representation does not rely solely on an absolute coordinate 
system. Therefore, a coherent world model can be 
constructed with noisy sensor data as long as the landmarks 
are properly recognized. Landmark recognition is central to 
the implementation of the proposed world model on a real 
robot. A neural network can easily recognize the natural 
landmarks selected. Two types of neural network (multi-
layer perceptron and learning vector quantisation) were 
trained and tested on a real robot for a natural landmark 
recognition task. 

Yüklə 0,71 Mb.

Dostları ilə paylaş:
1   ...   5   6   7   8   9   10   11   12   13




Verilənlər bazası müəlliflik hüququ ilə müdafiə olunur ©azkurs.org 2024
rəhbərliyinə müraciət

gir | qeydiyyatdan keç
    Ana səhifə


yükləyin