|
ResearchTelecommunication & Internet PolicyInformation Diffusion Model of Mobile Telecommunication Networks in Portugal from Social Network PerspectiveMore information will be added shortly. Intellectual Property Rights (IPR) & TelecommunicationThe Contemporary Intellectual Property Rights Arena on Competing Technologies Towards 4G, Master Thesis, Information Networking Institute, Carnegie Mellon University, 2009
Today knowledge has become pivotal resource for the modern economy. The need to protect innovative efforts thus is critically desired. Intellectual Property Rights (IPR) provided by the governmental agencies encourage innovators to disclose their valuable inventions to the public, in return for the temporarily granted legal rights to exclude other from imitation. As the surge of patenting activities appeared in recent 20 years,the current practice of patent system reveals to be controversial, especially with regard to the Information and Communication Technology (ICT) industry. The situation is even worse within the context of ICT standardization where IPR protection, standard setting and competition potentially conflict with each other. In this book, we conduct comprehensive review considering challenges brought by IPR system and convoluted interaction underlying the IPR, competition and standardization towards 4G. We then build up the evaluation framework which takes care of these complex relationships and mixed effects by means of IPR analysis. Finally we present the empirical evaluation result on Long Time Evolution , which will be the major path evolved from 3G towards 4G.
Computer ScienceContent-based Image RetrievalEfficient and Robust Two Stage Feature Fusion for
Image Classification, with Prof. Sofia Tsekeridou, Technical Report, Athens Information Technology, 2008
In this report we first describe the four general image classification paradigms in the real
world problems and then propose our two stage feature fusion framework accordingly. To
balance the tradeoff between the precision and efficiency while also pertaining the robustness,
our proposed methodology that uses subspace learning in early fusion while combining
different classifier techniques and their results is introduced in detail. Then we will extract
several features working on the VOC data set and widely used classification algorithms during
experimentation to assess the proposed methodology performance. The results indicate our
framework performs better when early fusion is deployed which is sufficiently justified.
Miscellaneous (Project Report, Course Report, Presentation, etc.)Presentation for Entrepreneurship class, 2011 Presentation for Multimedia class, 2008 Report for Policies of Wireless Systems and Internet, 2007 |