Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn about AT&T AutoClassify, an innovative end-to-end system developed collaboratively between AT&T's Chief Data Office and Databricks professional services for automatic multi-head binary classification from unlabeled text data. Discover how this novel approach automates the traditionally challenging process of creating labeled datasets and training multi-head binary classifiers with minimal human intervention, starting only from a corpus of unlabeled text and a list of desired labels. Explore the advanced natural language processing techniques used to automatically mine relevant examples from raw text, fine-tune embedding models, and train individual classifier heads for multiple true/false labels. Understand how this solution can reduce LLM classification costs by 1,000x, making it an exceptionally efficient solution in terms of operational costs, while producing a highly optimized and low-cost model servable in Databricks that can process raw text and generate multiple binary classifications. Examine a practical use case involving call transcripts to see the system in action, presented by Colton Peltier, Staff Data Scientist at Databricks, and Hien Lam, Senior Data Scientist at AT&T.
