Thalia is a recent graduate in Biostatistics and Economics from Simmons University. She is driven by intersection of information, engineering and philosophy that is Machine Intelligence. This summer she has been working with iD tech Coding Academy, teaching Machine Learning and Deep Neural Networks.
Performance Boosting Methods in Semi-Supervised NLP Neural Net Environments
My talk today will cover the use of AI/ML in natural language processing (NLP). The field of text analytics broadly refers to the component of artificial intelligence to understand human language, a topic made difficult by the inconsistencies in human communication. This talk with utilize a set of statistical techniques for identifying and grouping large text data sets, and how much machine models can be gained by using a preprocessed data which exhibits both semantic and syntactic similarity. This talk will be accessible without knowing the trends and research within the field today. It begins with a brief review of the rules and patterns and compares a logistic regression in percentages of accuracy to a reorganized model. I include some original research within vector-based recurrent nets and the leap in NLP performance that followed the representation of words as continuous vectors, which I will argue is the technique that will allow the field to grow the most effectively in the coming years.