The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) Generating SQL Database Queries from Natural Language with Yanshuai Cao - #519
17 snips
Sep 16, 2021 
Yanshuai Cao, a Senior Research Team Lead at Borealis AI, discusses his groundbreaking work on Turing, an engine transforming natural language into SQL queries. He compares it with OpenAI's Codex, highlighting the unique challenges of SQL generation. The conversation reveals insights into the crucial role of reasoning and common sense in accurate query creation. They also tackle complexities in multilingual datasets, data augmentation, and the ongoing quest for model explainability, shedding light on fascinating advancements in AI technology.
AI Snips
Chapters
Transcript
Episode notes
Adversarial Robustness Research
- Yanshuai Cao discussed his PhD work on adversarial robustness, where imperceptible changes to images altered their classification.
- Surprisingly, they could change features of an image to resemble another image entirely, like a car.
Model vs. Human Representation
- Adversarial robustness issues highlight the difference between how models and humans represent data.
- Models rely on shortcuts and spurious correlations, unlike human understanding and reasoning.
Turing's Purpose
- Turing allows non-technical users to interact with structured datasets and gain insights.
- It addresses the challenge of converting ambiguous natural language into unambiguous SQL queries.