Topic: Word Embedding
Speaker: Maryam Tavakoli Hosseinabadi
Abstract: Human natural language is based on semantic world of “words”, while machine learning algorithms are more comfortable with world of quantity and numbers. Many tricks and methods have been proposed to extract numeric features from alphabetic corpus, however it was always challenging to keep the spirit and meaning of the words in transiting from alphabetic world to the numeric one. Word embedding is one of the on going research areas that actively improves these mapping techniques to make these two worlds parallel and more similar. This seminar is not on a specific research work, but a brief and preliminary overview on word embedding concepts and current techniques and applications.