Abstracts of my Recent Publications |
1. Rule Based Personalization of Educational Videos : (Under review at the International Journal on E-Learning)
Abstract:
The amount of information available to the students in the form of texts, slides and videos is enormous. Hence the retrieval of needful content from the plethora of materials available becomes a Herculean task. Just like the internet, the problem with E-learning systems is information overload, and not the lack of it. Also, a dynamic E-learning system that can adapt to the varied user preferences is essential. Hence, for the system to be efficient, summarization and personalization of educational materials based on their semantic content is important. This paper addresses the need for basic rule sets in E-Learning systems that can simplify and automate the process of personalization of educational material. Different kinds of educational videos were investigated and simple generic rules that can yield the desired video summaries were developed.
2. Intelligent Routing in Dynamic Networks (To be sent for review soon)
Abstract:
As
the need for expansion of computer networking capabilities increase, the require- ment of an adaptive and reliable network routing solution becomes more prominent. Ef- fective dynamic routing protocols should take into consideration various factors like link speed, bandwidth, network traffic and topology and should readily adapt to changes in these factors. Though routing algorithms such as the Bellman-Ford and
This paper details the application of a variant of the Confidence-Based Q-Routing to IP routing in a dynamic networks. A classic example in which dynamic routing is imperative is in the establishment of Virtual Private Networks. A virtual private network (VPN) is the extension of a private network that encompasses links across shared or public net- works like the Internet. We analyze the effectiveness of Q-Routing and explain the significance of such a learning algorithm in the routing domain. Furthermore, we propose a variant of the Confidence Based Q-Routing and extend its application to Virtual Private Networks.
Keywords: Reinforcement Learning, Virtual Private Networks, Q-Learning, Q-Routing, Confidence Based
Q-Routing.