Automatic Text Summarization using Word Embeddings

As the amount of textual data on the internet continues to grow at a breathtaking pace, the need for a means to condense this data into smaller packets of more readily analysable/consumable information is obvious. To spend man-hours on manually performing this task would be counter-intuitive. A means of automatically producing summaries of texts using word embeddings is explored here. Word embedding refers to a set of techniques which map words from a textual corpus onto a numeric vector space, as per their context, thereby capturing an abstract sense of ‘word meaning’ in the vector space. The goal of this project is to usefully apply the concept of Word Embeddings to the task of Automatic Text Summarization, and to review existing techniques.