Databases Reference
In-Depth Information
8.4 Adaptive Recognition: A Different Perspective
In this chapter, we have presented a series of discussions on and exam-
ples of distributed pattern recognition (DPR) applications in coarse-grained
computational networks. Network granularity is important when considering
different types of Internet-scale data and applications. To achieve a scalable
recognition scheme, we believe that such applications must be adaptive to
different levels of network granularity. Recognition in the context of present
time is not limited to complex data analysis, which runs on high-performance
machines. Rather, we see the rapid development of lightweight devices that
can perform complex operations, e.g., sensors, mobile phones, and any other
wearable devices.
We have demonstrated the capabilities of a purely DPR scheme, such as the
DHGN, to address two different application perspectives in a coarse-grained
network, i.e., data recognition and management. Analyzing large-scale and
complex data, such as images, is di
cult without the availability of scalable
recognition schemes that work in computational networks. The criticality of
the information required, e.g., a medical analysis, adds to the importance of
having a fast recognition scheme that can scale up with large amounts of data.
In addition to image recognition, distributed data management is considered
to be a promising avenue for Internet-scale and distributed pattern recognition
applications. Using a data association approach, it is believed that the DPR
scheme can enhance the performance of existing data access schemes, such as
MapReduce, which has been proven effective in the cloud environment.
In the next chapter, we will consider applications of Internet-scale pattern
recognition in a different form of network, i.e., the fine-grained network.
