Currently, I am pursuing a PhD in Computational Neuroscience, Spiking Neural Networks and Machine Learning at University of York, UK. I hold a master’s degree in Digital Signal Processing specialized in adaptive signal analysis for EEG and EMG datasets from York and undergraduate in Electronic Engineering from Istanbul Technical University, Turkey.
My research during my PhD is currently based on SNN (Spiking Neural Network) and ML (Machine Learning). Also, I have started to work as consultant for a couple of start-ups on various projects related to matching, prediction and recommendation.
On the ML side, I have important experience on Reinforcement learning algorithms such as Temporal difference learning, Q-learning, Monte Carlo Method, SARSA, Dopamine modulated neural mechanisms and Knowledge-based Reinforcement Learning.I have extensive experience developing in Embedded C and Python. Projects which I have involved related with AI (Artificial Intelligence), ML (Machine Learning), SNN (Spiking Neural Network) will be demonstrated over this page.
The aim is developing efficient algorithms and machine learning techniques for better back-end web performance. The system itself is more related with prediction, recommender systems, data mining, and information retrieval.
Here I am interested in 1. applying statistical methods and machine learning to client data. 2. solve and prototype solution for specific problems. 3. helping in personal recommendation strategy for doctors and clinicians about published papers. Problem: Applying different machinery techniques in order to find ranking of papers for individual client that may help improving site’s searching and recommendation performance.
Not only the retrieved data directly from the HTML source also we extract features visually by using OCR (Optical Character Recognition). We use the open source Tesseract OCR engine, which was originally developed at HP and now primarily at Google. Each individual element is automatically detected, then we draw frame(s) around the data/element that we need. Aside from scraping all the possible texts in those detected frames, the position of texts and other text properties such as font size and colour are also captured. Based on the method, we label the data with proper names.
I will demonstrate more details about those projects here. Majority of those projects are implemented on Python language.