Virtual MLOps and Kubeflow Meetup - July 7

Zoom Link

*https://us06web.zoom.us/webinar/register/WN_xpli1UEoSjG3Bm69bepoLQ*

Agenda

• Welcome, Announcements & Housekeeping
• Talk #1: Intro to Kubeflow -- Jimmy Guerrero
• Talk #2: FiftyOne: the open-source tool for building high-quality datasets and computer vision models -- Brian Moore
• Talk #3: Jet Energy Corrections with GNN Regression using Kubeflow at CERN -- Daniel Holmberg & Dejan Golubovic

Intro to Kubeflow
Co-organizer, Jimmy Guerrero will give a 10-min, broad overview of the open source Kubeflow MLOps platform. We'll cover architecture, components, distributions and installation options.

FiftyOne: the open-source tool for building high-quality datasets and computer vision models

Nothing hinders the success of machine learning systems more than poor quality data. And without the right tools, improving a model can be time-consuming and inefficient.

In this talk, we'll do a brief overview and technical demo of FiftyOne, an emerging open source tool that provides the building blocks for dataset analysis and integrates with complementary solutions across the ML stack to solve the biggest existential threat to any ML project: dataset quality. Tens of thousands of engineers and scientists use FiftyOne everyday to get hands-on with their data, visualize complex labels, evaluate models, explore scenarios, identify failure modes, improve annotations, and much more.

Brian Moore is CTO/Co-founder of Voxel51, an AI software company that enables machine learning and computer vision scientists to rapidly curate and experiment with their datasets in order to build higher performing machine learning systems.

Jet Energy Corrections with GNN Regression using Kubeflow at CERN

The Large Hadron Collider is the world’s largest particle accelerator measuring 27 km in circumference. It accelerates beams of particles in opposite directions almost to the speed of light before making them collide. The particles emerging from the collisions are then measured in large detectors such as the Compact Muon Solenoid. An especially important object of study are so-called jets composed of multiple particles shooting out in the same direction from the collision point. Data-driven methods are used to correct the energy values for these jets, and what we’ll present here is the utilization of Kubeflow to enable state-of-the-art graph neural network based corrections. Kubeflow’s pipeline component allows us to define our machine learning workflow in a well-structured and reproducible manner, and its built-in training operators are used to scale up the training with ease. This work is expected to pave the way for future adoption of Kubeflow among the physics community at CERN.

Daniel Holberg is a technical student at CERN investigating deep learning applications for the CMS experiment. Dejan Golubović is a CERN software engineer with experience in machine learning.