+ 7:30 - 8:20 AM // Registration + Coffee
Grab your morning essentials + get ready for opening remarks!
+ 8:30 - 9:15 AM // Adam wenchel, Arthur AI
Adam Wenchel is the CEO of Arthur AI, the leader in enabling enterprise-grade AI. Prior to ArthurAI, Adam founded Capital One’s Center for Machine Learning (C4ML) and led transformative ML projects across the enterprise. Adam started his career as an AI researcher at the Defense Advanced Research Projects Agency (DARPA) developing automated planning technologies and has been active in the field of ML/AI for 20 years. Adam is a regular speaker at national conferences on ML and has been featured in the Wall Street Journal, Forbes and several other national publications on deploying AI responsibly.
+ 9:30 - 10:15 AM // Ranju Das, Amazon AI
Ranju Das is part of the technical leadership team at Amazon Web Services, pulling from over a decade of expertise in distributed systems, architecture, web-scale analytics, big data, machine learning and high performance computing to help customers bring their ideas to life through technology. Since joining Amazon in early 2013, he has played a role in introducing significant new features and products to retail customers like Amazon Drive, and Amazon Prime Photos. Ranju has led the development of the Amazon Rekognition Service from its inception. Prior to Amazon, Ranju played a key role in the delivery of Nook Tablet, led the big data and database development for Barnes & Noble, and was founding member of two startups.
+ 10:30 - 11:20 AM // Breakout V.1
// IMPACT OF DATA SCIENCE ON BUSINESS -
Speakers: Rob Reynods (Markel) + Mayur Rajadhyaksha (Altria) + Arka Chakraborty (UVA) | Moderator - John Harris (Boxelder Analytics)
Data is skyrocketing in its importance and value creation. In this panel discussion, we will explore how Altria, Markel and UVA are using data, machine learning and artificial intelligence to create business impact.
// TEXT REPRESENTATIONS FOR DEEP LEARNING -
Speaker: Zachary Brown (S&P Global)
In this talk, we'll discuss approaches to creating useful vector representations for text using deep learning frameworks. We'll begin talking about the difficulties with traditional count-based local representation, then discuss how dense representations are calculated, highlighting several state of the art techniques for creating vector text representations.
// MACHINE LEARNING AT THE EDGE: CHALLENGES AND CONSIDERATIONS -
Speaker: Miriam Friedel (Metis Machine)
At Metis Machine, we believe that we are at a renaissance, similar to what we saw in 2008 when the Apple App Store opened. Today, thanks to a host of new technologies, we have the opportunity to create ML-powered apps that change what is possible. However, in spite of the incredible opportunities that are now available, successfully deploying machine learning to the edge has a unique set of challenges and considerations.
// ACCELERATING DATA SCIENCE WORKFLOWS WITH RAPIDS -
Speaker: May Casterline (NVIDIA)
The open source RAPIDS project allows data scientists to GPU-accelerate their data science and data analytics applications from beginning to end, creating possibilities for drastic performance gains and techniques not available through traditional CPU-only workflows. This talk will provide an introduction to the open source project, examples of where GPU-acceleration has had an impact in traditional data science workflows, and resources available to get started.
+ 11:35 AM - 12:25 PM // Breakout V.2
// HOW TO SCALE A DATA ENGINEERING TEAM -
Interview with Paul Hurlocker (Capital One) | Moderator - Patrick Harrison (S&P Global)
Paul Hurlocker, VP in Capital One's Center for Machine Learning, and Patrick Harrison, Director of Data Science for S&P Global Market Intelligence, will share their real-world experiences building high performing machine learning teams and successfully positioning them in an organization. The interview format will cover reasons to build the capability, the types of roles and skillsets required, approaches to developing and recruiting talent, and some of the business and technical precursors that improve the probability of success.
// TIME SERIES, R, + CHIPOTLE -
Speaker: Pablo Ormachea (West Creek Financial)
Do you want to learn how to measure seasonality? Or how to build effective forecasts for time series data? Join Pablo, Richmond-based West Creek Financial's VP of Data Science, as he walks you through the time series data ecosystem in R. We'll quickly build valid models to forecast stock performance in a few different ways - from the basic to the cutting-edge.
// DESIGNING DATA PIPELINES FOR MACHINE LEARNING APPLICATIONS -
Speaker: Alexis Seigneurin
Over the years, training a Machine Learning model has become fairly easy. It is time to take this to the next step: using your model in a streaming application. In this talk, we will see the pitfalls and the best practices for you to create data pipelines with Kafka and Machine Learning models.
// A PRACTICAL GUIDE TO FEATURE REDUCTION TECHNIQUES -
Speaker: Vishal Patel (Derive)
Most machine learning applications involve datasets with high dimensionality. In most cases, the intrinsic dimensionality is much smaller than the observed dimensionality of the data, and it becomes imperative to eliminate unavailing and redundant features before performing the core analysis. This talk provides a step-by-step overview and demonstration of several feature reduction techniques pertaining to the pre-processing of data for supervised learning problems. Attendees should have some basic level of understanding of data wrangling and supervised learning.
+ 12:30 - 1:30 PM // Lunch Break
A boxed lunch will be provided by White House Catering
+ 1:30 - 2:20 PM // Breakout V.3
// DATA SCIENCE + CUSTOMER IMPACT -
Speakers: Ryan Ehrensberger (UNOS) + Nick Anderson (CarMax) + Ankit Mathur (RoundTrip)| Moderator - Ben Harden (CapTech)
// LEARNING TO RANK SEARCH -
Speaker: Elizabeth Haubert (OpenSource Connections)
// SCALABLE DATA STORAGE + PROCESSING -
Speaker: Todd Nemanich (APIvista)
// UNDERSTANDING DATA THROUGH VISUALIZATION -
Speakers: Patsy Daniel (Compare), Eric Jenvey (McGuireWoods), Tessa McKenzie (Impact Makers) | Moderator: Spencer Hansen (Simple Thread)
Open discussion around telling stories with data and considering the user experience when visualizing data.
+ 2:35 - 3:25 PM // Breakout V.4
// SO, YOU WANT TO BE A DATA SCIENTIST? -
Speakers: Jackie Goldschmidt (APIvista) + Paul Brooks (Virginia Commonwealth University) + Nick Anderson (CarMax) | Moderator - Joanna Bergeron (CapTech)
According to LinkedIn and Glassdoor, Data Science jobs grew by 56% last year making it the fastest growing job category in the US. Panelists from CarMax, VCU and APIvista will provide advice on how to get started in in this career path describing the different kinds of jobs available in Data Science as well as the skills and training required for each.
// COMPUTER VISION + OBJECT DETECTION -
Speaker: Myles Baker, Databrick
Innovations like facial recognition, autonomous vehicles, gesture recognition, and augmented reality are flourishing as digital interfaces that capture and use image data pervade our society, particularly in the space of consumer electronic devices. This talk introduces the technical foundation of computer vision by way of an in-depth demo of video object detection using the YOLO real-time object detection system. The talk will close with a brief review of the ethical debate in computer vision.
// AUTOMATED MACHINE LEARNING: HOW IT IS CHANGING DATA SCIENCE -
Speaker: Chandler McCann (DataRobot)
// CAN A MACHINE BE RACIST OR SEXIST -
Speaker: Renee Teate (HelioCampus)
Is it possible for a machine learning model to be socially biased? Does leaving a decision up to a logical computer remove human biases? How do predictive models work? At what steps in the model development process could social bias be introduced? In this talk, we'll explore these questions and more to discuss concerns that business leaders and data scientists should be aware of regarding machine learning and AI, and their impact on individuals and communities.
+ 3:40 - 4:30 PM // Chris Peterson, Capital One
Chris Peterson, Managing Vice President, leads the Data Science organization in US Card with responsibility for developing and maintaining the core underwriting risk and valuation model systems, developing data and model scoring platforms, and defining Card’s data strategy. He is the accountable executive overseeing relationships with the primary credit bureaus. Chris also leads the Data Science recruiting process across the enterprise.
Chris has been with Capital One since 2006, with prior roles in Card and the Model Risk Office, overseeing models in Capital One Auto Finance, Mortgage, International Card, Anti-Money Laundering, Retail and Commercial Bank. Prior to joining Capital One, Chris worked at Intel, 3M, and the National Security Agency.
+ HAPPY HOUR //
We reached the SUMMIT + now it's time to celebrate! Enjoy an after conference happy hour!