ABD SOCIETY & UNC CHARLOTTE’S

DATA ANALYTICS SPEAKER SERIES

[WEBINAR]

formal and probabilistic approaches to programmatically understanding the natural language

Thurs, MAY 21st, 2020

12pm - 1pm

WATCH RECORDING

MEET our speaker

Dima Korolev | LinkedIn

Head of Machine Learning at FriendlyData, acq’d by ServiceNow

Dima started with NLP and machine learning back when it was called Information Retrieval in 2005, when the bleeding edge for text categorization was to use SVM over log-scaled TF*IDF bags of words. After graduating, several patents and several years at Google later, he switched focus and became an entrepreneur and consultant. His most recent gig was getting acquired as part of a small team FriendlyData Inc., a company that has developed and productionized the high-accuracy English-to-SQL solution based on strict grammars to parse the queries in natural languages. While he doesn't consider himself an NLP expert, he enjoys speaking about the different approaches for the machines to understand the natural language, their strengths and weakness, and the future that lies ahead.

PRESENTATION TOPIC

Formal and probabilistic approaches to programmatically understanding the natural language

The field of NLP is surging these days, and it's easy to get lost in the whitepapers and presentations that are frequently published by big companies. At the same time, while the novel techniques require the training corpora and the machine resources that greatly exceed what the small teams have access to, there exist relatively simple and straightforward solutions for most niche problems. Employing these solutions, in fact, has shifted quite a bit in the recent years, and is now largely in the domain of engineering than in the domain of science, and Dima would like to speak about the status quo in the industry, as well as what low hanging fruit is left to be picked.