Two Talks on Deep Learning at Metis

NOTE: Please sign in at the Metis lobby entrance before entering. Thank you!

With the support of Intel, BeMyApp, and Metis, we are featuring two talks on deep learning. Food and drink will also be provided.

Talk 1 - Using Deep Learning for for Image Segmentation in Keras

Gavin Patient attempted the Kaggle Carvana competition for his capstone project. The problem was to crop cars out of images for marketing purposes, though it was also a great excuse to read some deep learning papers on image segmentation and mess around with architectures in Keras. He is here to talk about the challenges of catching up with research and applying it to real world problems without a PhD to back it up.

About the Speaker

Gavin Patient comes to Metis with a BS in both Mathematics and Physics from the University of Michigan. He has experience working inside research labs and testing software. Drawn to data science by the recent advances in machine learning, Gavin has become passionate about understanding the limits of the field and its applications. His final project utilizes deep learning methods to crop cars out of images.

Talk 2 - Using Deep Learning in Ad Photo Ranking

Marco Zamora regularly shoots and post-produces food photography for his family-owned small retail business's social media ads. He wanted to explore how a neural network could help streamline that process for him to publish more and higher quality ads with less effort.

The resulting classifier cut down manual processing time by about 80% for the first part of image post-production, using transfer learning on top of a VGG 16 ImageNet convolutional neural network built with Python and Keras on a GPU-equipped Google Compute instance.

In his talk he will go over the general constraints and design decisions, his experience wrangling with Keras, AWS, and Google's cloud services, and possible future steps in this problem domain.

About the Speaker

Marco Zamora is driven by investigating how things work in order to make them work for you. He has an academic and personal background in math and computer science, and a professional trajectory in data and analytics with experience in application lifecycle management for information technology and financial services.

Prior to joining Metis, he was a manager in Avanade, a subsidiary of Accenture, delivering solutions for large volume transactional data analysis and processing, and solutions for digital experiences. He also owns a full-service bakery with his wife. For his final project he wants to automate his business advertising post-production workflow with neural networks for image processing, to help him attract more customers with higher frequency, high-quality advertising.

About our Sponsors

Food and drinks will be available at the event, courtesy of BeMyApp.

Venue and speaker, courtesy of Metis.

Intel also created a mini contest for all the participants. If you have a ML project and want to showcase it, share it, or collaborate it with others, submit it to DevMesh. Everyone who submits a project to DevMesh will get remote access to Machine Learning Servers. On top of that, best projects will be selected and each winner will receive a $50 gift card.

Instructions to join DevMesh:

1/ Create a new account at devmesh.intel.com (http://devmesh.intel.com/)

2/ Join your dedicated group - Artificial Intelligence West Coast (https://devmesh.intel.com/groups/449)

3/ To submit a project, click on “add a project”

*when submitting your project, make sure to select “Artificial Intelligence West Coast (https://devmesh.intel.com/groups/449)” as your group.

To receive invitations for Intel webinars, news and tools for Machine Learning and Deep Learning, register on this link (https://plan.seek.intel.com/us_en_software_registration-form_DeveloperMeetupFollow-up_html?registration_source=florida_meetup).