Special Presentation Night - The Three Dimensions of Scalable Machine Learning

Thank you to Akamai for both hosting and paying for food and drink for this last-minute event!

We are joined by Reza Zadeh (Standford University, Databricks) for a special talk on machine learning. Reza built the Machine Learning Algorithms behind Twitter's who-to-follow system, the first product to use Machine Learning at Twitter. He is also the initial creator of the Linear Algebra Package in Apache Spark.

Join us for this special event! We will be starting at 6:40 sharp. Please bring your IDs to check-in at the front desk.

---

Speaker: Reza Zadeh

Title: The Three Dimensions of Scalable Machine Learning

Abstract:
Machine Learning can be scaled in three different ways: 1) Data Parallelism for large number of examples, 2) Model parallelism for large models, 3) Lots of little models for hyper parameter tuning.

In this talk, we will begin with an overview of data flow computing models and use them to scale machine learning algorithms, with a focus on Apache Spark and MLlib. We consider scaling linear models, matrix factorization, and others.

Bio: http://stanford.edu/~rezab/bio.txt

---

For those driving to the event, there is parking nearby.

Parking Garage: Parking is available at the Cambridge West Garage in Cambridge Center off Ames Street. This is a public/paid garage -- you can check your options here: http://en.parkopedia.com/parking/garage/cambridge_center_west_garage/02142/cambridge/