Dr Oluwatoyin Enikuomehin
My research span over a number of areas, including statistical machine learning, text and natural language processing including BioMedical NLP, information retrieval, information theory and recently computer vision. Mainly, my research on Information Retrieval (IR) covers all back-end aspects, from crawling and indexing to ranking, evaluation, efficiency/scalability and analytics, and its extension to Computer Vision. Essentially, I have tailored some of my work towards query/document representation, ranking extensions (e.g. term dependence, discourse analysis), and domain-specific retrieval (especially in rare medical diseases, Medical diagnosis, including Magnetic Resonance Imaging data analysis, scholarly (big) data, recommender systems and digital learning systems).I have also been working on the intersection of IR and Natural Language Processing (NLP). These two areas share a lot of common ground, in terms of the statistical means employed to process information, but also challenges, in the need to scale their respective approaches to noisier datasets, or to transpose supervised lab-based methods to naturalistic real life situ-environments, as exemplified in my recent paper titled: Computational analysis of Ondo Accent in the pronunciation of English word. My research aims to explore how much and under which conditions we can improve the IR systems through more linguistically-oriented information processing components. Sometimes, my research leads me into dwelling in some NLP themes (without an IR angle), e.g. plagiarism detection/authorship attribution, summarization, sentiment/opinion mining. In addition, I conduct research in data mining using information available from online social networks and search engine query logs for healthcare purposes. With a solid background in mathematics of computation, specifically mathematical logic, computability theory, and computational complexity theory, with the aid of experts in medical imaging, I have been working to develop tools for automated medical image analysis provided as a web service. For instance, my recent work focuses on Brain Atrophy in African stroke survivor. I have also worked on the modeling of cognition for shared intelligence (Ongoing). I focus on presenting the most appropriate information that will satisfy the information seeking task of a User. The purpose of my research in IR is to provide further solutions to the information retrieval problem using some computer-based mathematical framework. This problem coined the IR problem becomes more complex as the dataset increases. The exponential growth of available information on databases, web, repositories etc has further compounded the problem. A way to understand how these vast data can be modeled is by studying deeply the Natural Language Processing techniques which tends toward semantic assistance to effective retrieval. I have diverse a language modeling approach such that I engage natural techniques in NL processing for IR. I have liaised over the course of my research with one of highly professional personnel in Quantum IR (Prof Keith van Rijsbergen) for language modeling. Future Direction I am currently undertaking some research in some core areas of Health Informatics especially Electronic Medical Record Modeling. My aim is to use Natural Language Processing tools to predict, detect, monitor and in extension manage Adverse Drug Effects(ADEs). As a Lead investigator in my computer vision team, we are currently looking at some computational methods of predicting the level of white matter hyperintensity for stroke survivors in Africa. It is my wish to extend the use of NLP in health domain. I am open to research collaboration in many other domain. Also, in the current time, I am deeply working on Sentiment Analysis and Opinion Mining across many social media platform.