This paper proposes a new approach for a hybrid connectionistHMM speech recognition system. The system consists of a multi-feature HMM-based recognition module using three different neural networks as multiple neural codebooks. Each neural network receives a different feature (i.e. cepstrum, delta cepstrum, and delta power) as input and generates a vector quantizer label obtained from the firin...

Although populations of amphibians are declining worldwide, there is no evidence that salamanders occupying small streams are experiencing enigmatic declines, and populations of these species seem stable. Theory predicts that dispersal through multiple pathways can stabilize populations, preventing extinction in habitat networks. However, empirical data to support this prediction are absent for...

iii ACKNOWLEDGEMENTS There are so many that have shaped this endeavor in so many ways, it is impossible to thank all of them. I apologize for any omissions, and I hope you understand you are in my mind though not in this section. First and foremost I would like to thank my parents for taking the courage to send their 18 year old child to United States. While I pursued my education at Tech, I kn...

Online monitoring of data streams poses a challenge in many data-centric applications, such as telecommunications networks, traffic management, trend-related analysis, webclick streams, intrusion detection, and sensor networks. Mining techniques employed in these applications have to be efficient in terms of space usage and per-item processing time while providing a high quality of answers to (...

Sequential pattern mining is the mining of data sequences for frequent sequential patterns with time sequence, which has a wide application. Data streams are streams of data that arrive at high speed. Due to the limitation of memory capacity and the need of real-time mining, the results of mining need to be updated in real time. Multiple data streams are the simultaneous arrival of a plurality ...

This paper presents a novel approach, called MODS, to build an accurate time evolving classifier from multiple one-class data streams learning time evolving classifier. Our proposed MODS approach works in two steps. In the first step, we first construct local one-class classifiers on the labeled positive examples from each sub-data stream respectively. We then collect the informative examples (...

The geometric complexity of stream networks has been a source of fascination for centuries. However, a comprehensive understanding of ramification--the mechanism of branching by which such networks grow--remains elusive. Here we show that streams incised by groundwater seepage branch at a characteristic angle of 2π/5 = 72°. Our theory represents streams as a collection of paths growing and bifu...