Learn about Cleanlab's innovative AI-powered tools for correcting errors in datasets during this 25-minute conference talk by Curtis Northcutt, CEO and co-founder of Cleanlab, at the 2022 Future of Data-Centric AI conference. Explore the problem of data quality in AI model training, understand the key ideas behind Cleanlab's approach, and discover the theory and formulation of Confident Learning methods. Gain insights into data-centric AI practices and various use cases for data cleansing. Watch a demonstration of Cleanlab Studio and understand how it can be applied to improve the quality of datasets across different domains.
Overview
Syllabus
Introduction
Agenda
The Problem
Labelers
Results
Key Idea
The Results
Theory
Formulation
Confident Learning Methods
DataCentric AI
Use Cases
Cleanlab Studio
Demo
Taught by
Snorkel AI
Related Courses
-
Automatically Fix Data Issues and Label Errors in Most ML Datasets - Cleanlab
-
Training AI with Humans
-
Operationalizing Organizational Knowledge with Data-Centric AI
-
Operationalizing Data-Centric AI - Practical Tools to Quickly Improve ML Datasets
-
Solving the Last Mile Problem of Foundation Models with Data-Centric AI
-
Why AI Needs New Data Benchmarks and Quality Metrics
