Computer vision on mobile device & Multi-stage ML for document understanding

Streaming Link:
https://www.meetup.com/SF-Big-Analytics/events/258514786/

Agenda
6:00 pm -- 6:30 pm light food + networkinig
6:30 pm -- 6:35 pm introduction
6:35 pm -- 7:25 pm Talk 1 (Facebook) + Q&A
7:20 pm -- 8:10 pm Tak 2 (Workday) + Q & A
8:30 pm Closing

Talk 1 "Efficient Deep Learning for Computer Vision on Mobile Devices"

In this talk, we will first talk about a few augmented reality applications that are based on computer vision in the family of Facebook products. Then, we will dive into talking about three deep learning advancements that has enabled the computer vision algorithms that power these applications, i.e., model compression, automatic architecture search, and Mask R-CNN2Go.

Speaker: Sam Tsai (Facebook)

Sam Tsai is an applied research scientist from the Mobile Vision team in Facebook where his is working on efficient deep learning computer vision solutions for mobile and embedded applications. Prior to Facebook, he worked in A9's visual search team that developed the recognition engine behind Amazon's camera search, and also in Nokia Research Labs building visual search solutions. He holds an PhD degree from Stanford and has authored more than 50 academic papers and 10 patents.

Talk 2: Document Understanding using Multi-stage Machine Learning

Workday is an enterprise SaaS company with products and services in the areas of HCM, Financials, Payroll, Analytics, Planning, etc. Our use of machine learning spans across all product areas. Today we will be talking about how we apply machine learning to automate document information extraction business processes which have been traditionally done manually or semi-manually. The techniques we have applied are deep learning based approaches in Computer Vision and Natural Language Processing. The results are competitive with major public cloud providers and have high rate of automation.

Speakers: Henry Zhang and Vivek Srivastava (Workday)

Henry Zhang is lead Machine Learning Scientist and Engineer in the machine learning team at Workday. His current focus is on machine learning technology that can be applied to multiple product use cases across the entirety of enterprise software space. He holds a B.S. and M.S. from University of Waterloo and University of Toronto in Computer Science.

Dr. Vivek Srivastava is the Machine Learning Product Manager at Workday. He has over 10 years of experience in building scalable products. Currently, he is responsible for roadmap creation, product evangelization and prioritization for the ML platform of Workday. Vivek has a PhD in Computer Engineering from Virginia Tech.