GraphLab - Machine Learning for Big Data in the Cloud

Explore a keynote presentation on GraphLab, a framework for machine learning on big data in cloud environments. Delve into the challenges of designing and implementing efficient parallel ML algorithms for handling large-scale data. Learn about GraphLab's ability to express asynchronous, dynamic graph computations crucial for state-of-the-art ML algorithms. Discover how GraphLab addresses parallelism challenges, including data distribution, optimized communication, and sequential consistency. Examine the framework's performance improvements over Hadoop in various large-scale tasks. Gain insights from Carlos Guestrin, a professor of Computer Science & Engineering at the University of Washington, as he discusses the growing importance of machine learning in industry and science, and the need for scalable, parallel ML algorithms to tackle increasingly complex tasks in the era of big data.

GraphLab: Machine Learning for Big Data in the Cloud—Carlos Guestrin (UW CSE)