Transfer Learning: Applications for natural language understanding

Abstract: Deep learning has been transformational in many aspects of AI and powers many business use cases. However, there are limits to what can be done with rote application of deep learning. One such limitation appears when you have a specific domain to understand, but all the available training data is for a different or more general domain, and acquiring the vast amount of data necessary to train a new system is infeasible. This is where transfer learning provides new opportunities — using the insights of one system to enrich another.

ML pioneer Andrew Ng has called transfer learning "the next driver of ML commercial success." Transfer learning makes powerful systems more reusable, and reduces the amount of training data, compute, and professional services needed. Is it ready for business deployment or is it still emerging technology? How is it used in business today? This talk focuses on language related use cases for customer service, search, question answer, self-help and consumer finance. We'll also have some fun with applications of transfer learning.

Bio: Dr. Catherine Havasi is the CEO and Co-Founder of Luminoso Technologies, an Artificial Intelligence (AI), natural language processing (NLP) company in Cambridge, MA. Luminoso was founded on nearly a decade of research at the MIT Media Lab on how NLP and machine learning could be applied to text analytics. For over 15 years she has been researching language and learning and was a research scientist in artificial intelligence and computational linguistics at the MIT Media Lab where she ran the Digital Intuition group. In the late 90s, she co-founded the Common Sense Computing Initiative, or ConceptNet, a big-data lexical resource used in over two thousand academic projects.

Open Data Science Conference