Fall

Using Deep Learning to Classify Pregnancy Test Results

World Health Organization estimates that there are approximately 36 million people with blindness. It's also estimated that around two thirds of this popu- lation is women, which is approximately 24 million. From discussion made with Bogazici University As- sistance and Education Center for Visually Impaired (GETEM), we have learned that young, visually im- paired women have no private way of taking a preg- nancy test and understanding the result. They ei- ther need to get outside help, be it a friend, family, or spouse, or they need to visit a hospital.

Short Text Classification with Novel Smoothing Kernels Based on Semantic Values of Terms

In literature, traditional classifiers are used in classification studies done on social network sites and semantic relations of words are ignored. They ignore the natural behaviors of natural languages like synonym, polysemy, multi-word expressions and latent-semantics.

BlockLearn

Since its introduction with Nakamoto’s famous Bitcoin paper, the concept of blockchain has been gaining attraction all over the world. While development of new technologies is becoming more of an issue, teaching how to use them in an effective way provides new opportunities and challenges. Blockchain technologies has their own understanding in the manner of learning and teaching. Therefore teaching blockchain concept and programming requires innovative approach in order to reach better learning experience.

Distributed Machine Learning Algorithms on Large Datasets

Machine learning algorithms are widely used in engineering and business.These algorithms scale and speed varies over different applications. Most of them relies on vast amount of data in limited amount of time and space. To handle these vast data we need distributed versions of the algorithms we have. In this work parallelization of machine learning algorithms were main problem focused on. The problem tried to parallelize is matrix factorization. We used stochastic gradient descent approach to minimize a linear loss function such as RMSE.

Effects of Obstcle in Axelrod's Culturel Model

There are a lot of different races, a lot of countries and communities of people in the world. it was thought that the small communities had taken under the large communities were assimilated and compared to the big societies. Each community maintains its unique characteristics. And they take property from other societies to the extent they allow. In some societies there are completely obstacles. These obstacles lead to the complete elimination of cultural, social, environmental or religious transitions.

Pet Tracker

Aim of the project is giving a chance to people in our campuses for getting information about pets in the campuses and making job of people who are in the animal house and organizations like Bupati etc. easier. The information is stored on a server and there is Bluetooth Low Energy devices on pets. People can see the information about nearby pets using the Android application. Also, with the help of animal houses and organizations like Bupati, the written data will be moved to the server and updated by voluntary vets and them.

XPath Language Parser For Validating Financial Documents

Turkey is in the process of changing from paper bills to electronic documents. These files carry/store crucial information and assumed valid under governmental issues, therefore they should be validated. In the business world, such documents are stored in UBL format as an XML file, and they are validated against a schematron file. Schematron assertions are written with XPath expressions, .NET environment does not support XPath 2.0 and above, however there is already version 3.1, and workarounds yield poor performance.

Evaluation Of Movie Recommendation Algorithms

Recommendation systems are gaining importance, especially with the development of new technologies and advanced machine learning algorithms. However, in the of case recommending items to users we have to be very careful about which dimensions to consider. We have to consider the domain of recommendation and also the purpose of the recommendation. Moreover, we have to come up with the best system, given the data we have, that is good at recommending items to the users.

Tool Development for Goal-Oriented Release Planning

User stories are useful for agile development methodologies for they are easy to learn and document. However, a list of user stories does not capture the underlying relations among the user stores, such as increase-customer-value relation between a functionality and its corresponding user interface. Goal models, on the other hand, quite capable of capturing several relations among requirements modelled as goals and help human experts answer questions such as 'how to achieve a requirement’ and ‘why does a certain requirement exist’.

Nanonetworking Testbed SW Platform Design

Nanonetworking is an emerging research field that focuses on communications between nanomachines. Nanoscale physics renders traditional electromagnetic communication mechanisms useless; therefore, novel approaches are required. Molecular Communications (MC), thanks to the help of the evolution of biological organisms, offers the most promising approach in Nanonetworking. Visit our Nanonetworking Research Group (NRG) website at http://nrg.boun.edu.tr for details.

Search form

Main Menu