Bachelor Thesis von Nicolas Thule

Abstract

Word Embeddings are a major trend in the Natural Language Processing (NLP) community. Recent years showed a phenomenal explosion of the literature corpus in the research field. Motivated by these indicators to clarify the meaning of this trend and studying the utilization of Word Embeddings and their different Use Cases, quickly a major drawback of the vast and steadily growing amount of literature in the research field revealed itself. The body of literature lacks of comprising introductory overviews of the methods and concepts Word Embeddings are researched for. Trying to fill this gap, this thesis additionally introduces a novel layer model as a first step to an overall tool helping beginners in the field to gain fundamental insights to methods and NLP tasks leveraging Word Embeddings.
This work conducted an analysis and classification approach on a  literature snapshot based on a greatly researched, discussed and adapted publication which is often mentioned as the initializing starting point of the advent of Word Embeddings. Evaluation of the literature snapshot built the basis for the developed layer model and shed some light on interesting insights, gathered throughout the developments process of this thesis. Presenting possible improvements for the layer model and future directions, the thesis concludes with a short summarization of the contributed contents.