Easy Deep Learning Text Classification With gobbli: J Nance & P Baumgartner, RTI

Every 2nd Tuesday of the month
LifeOmic
3800 Paramount Pkwy Suite #150 · Morrisville, NC
How to find us
Not far off the intersection of I-40 and 540, in the Perimeter Park area. Parking freely available.
Details
Easy Deep Learning Text Classification With gobbli – Jason Nance and Peter Baumgartner (RTI)
Classic machine learning models for text classification work best when you have large, labeled datasets to learn from. But what happens when you have a small dataset of text that you need to classify? Transfer learning models have unlocked new possibilities for applied natural language processing, enabling strong results on smaller datasets. However, the extreme complexity of these models makes them difficult to apply to problems in social science and survey research.
gobbli, a new open-source Python library developed by RTI International data scientists Jason Nance and Peter Baumgartner, is a friendly face in the murky depths of deep learning for text classification. It provides a uniform interface for multiple state-of-the-art models and other helpful utilities, such as data augmentation and experimentation. This talk will cover the improvements available through transfer learning, showcase the new library's features, and cover actual and potential use cases.
https://medium.com/rti-cds/introducing-gobbli-b625c0a0adfe
https://pypi.org/project/gobbli/
• Important to know
Ground floor, first door on the right as you come in the main entrance of 3800 Paramount Pkwy, Morrisville.