Spring

Cellular automata (CA) consists of a simple and well-formalized model of massively parallel computing, known to be capable of universal computing. CA has rich information processing capabilities because of their parallel behaviour; however, defining their power limitations is not easy. It is a useful approach to classifying the computational capacity of CA to examine their complexity classes.

Throughput of any process is the main indicator of the quality of a system. Whether it is the speed of the storage unit, random access memory or internet connection; throughput is the rst system performance metric to consider when choosing a system design. Since the creation of the internet, most of the research and development was focused on increasing the speed and reliability of the connection. Internet service providers, access point manufacturers and many competed to sell best product which can provide fastest network connection.

There are some neurologic related diseases that might be seen in the same person and whether there are comman factors that lead to the situation or not has been the subject of interest in biomedical science. Here, by using text mining with articles, our aim is to propose a connection between diseases in gene base. In this project, spesifically, we focussed on the connection between Autism Spectrum Disorder & Epilepsy Disorder and derivatives because they often occur together. We have used PubMed to retrieve related articles and created gene-gene graphs with them.

We propose to develop a machine learning system which can predict which home users of an Internet Service Provider will call to within one week. Training data for the system will be provided via per-minute data gathered from modems by the Lifemote, labeled both manually and by the ISP call center via the Lifemote user interface. Roughly 3 percent of users call their ISP to complain each month. The goal of the project is identify the 10 percent of the population most likely to call with a false negative rate of less than 20 percent.

A new neural network architecture to extract key frames is proposed. We combine ideas from both Learning to Explain: An Information-Theoretic Perspective on Model Interpretation, which is a work on instance-wise feature selection, and Concrete Autoencoders for Differentiable Feature Selection and Reconstruction, which is our previous work on unsupervised feature selection. Experiments are conducted on the BosphorusSign dataset, classification and reconstruction errors of the selected key frames are measured.