Academic and industry research in Natural Language Processing (NLP) has progressed at an accelerating pace over the last several years.
Members of the Python community have been hard at work moving cutting-edge research out of papers and into open source "batteries included" software libraries that can be applied to practical problems.
Our speaker for Python Charlottesville on this evening is Patrick Harrison. Patrick is Associate Director of Data Science at S&P Global Market Intelligence, and is an active practitioner of NLP with Python. He is also co-author of an upcoming O'Reilly book entitled "Deep Learning with Text" (Amazon pre-order link: http://amzn.to/2AkjOFw) with Matthew Honnibal, who is the creator of spaCy, an actively-developed NLP library for Python.
Patrick will present how to approach Modern NLP with Python. For those of you used to toy examples with NLTK, this will be a refreshing upgrade to your toolset, as NLTK is now widely considered to be a toy/teaching library, and is being rapidly replaced by more modern (more effective and more performant) techniques, powered by other open source libraries in the Python community.
Note 1: Before Patrick's talk, Andrew Montalenti, co-founder & CTO of Parse.ly, will present a 10-minute lightning talk on his experience with NLP using NLTK in the past, and the results of a recent bake-off he performed of spaCy's named entity extractor against a couple of commercial cloud NLP APIs, using a real-world English news dataset.
Note 2: We moved the start time of this meetup up slightly to 5:30pm. The lightning talk will happen 5:50pm-6pm, and then the main presentation will start promptly at 6pm. First 20 minutes can be used for food, drinks, mingling. Feel free to show up a little late if you can only make it at or after 6pm.