Loading
  • Online: Pitfalls and Challenges of ML-Powered Applications

    Building machine learning applications can be complex. Choosing the right ML approach for a given feature, analyzing model errors and data quality issues, and validating model results to guarantee product quality are all challenging problems that are at the core of the ML building process. Join us in August to learn about some of the pitfalls in deploying ML applications. Agenda ------------------------------------------------- 12:00 PM -- Greetings 12:05 PM -- Pitfalls and Challenges of ML-Powered Applications - Emmanuel Ameisen Location ------------------------------------------------- Zoom and YouTube Streaming A link will be sent out prior to the event. Please note that Zoom is capped at 100, so if you do not get into the Zoom, you will be able to watch via YouTube. Talks ------------------------------------------------- Pitfalls and Challenges of ML-Powered Applications As a field, we often hear about success stories. This is true in research, where a publishing incentive can pressure authors to focus on consistently exceeding state of the art results. It is also true in industry, where companies attempt to attract engineering talent by describing how impressive their production ML systems are. However, every practitioner here knows that in engineering and in ML, the road to success is paved with failures. The field of ML in production is new, and so has a lack of cautionary tales of things that can go wrong with models. This talk will try to help correct that. We will discuss challenges such as performance mismatch between offline training and online inference, feature generation and data leakage, and adequate roadmap planning for ML. Speaker ------------------------------------------------- Emmanuel Ameisen is a machine learning engineer at Stripe. He is the author of the book "Building Machine Learning Powered Applications" Previously, he led Insight Data Science’s AI program, directing more than a hundred machine learning projects. Before that, he implemented and deployed predictive analytics and machine learning solutions for Local Motion and Zipcar. Emmanuel holds graduate degrees in artificial intelligence, computer engineering, and management from three of France’s top schools. Follow on Twitter @mlpowered. Resources ------------------------------------------------- Building Machine Learning Powered Applications: Going from Idea to Product Learn the skills necessary to design, build, and deploy applications powered by machine learning (ML). Through the course of this hands-on book, you’ll build an example ML-driven application from initial idea to deployed product. Data scientists, software engineers, and product managers—including experienced practitioners and novices alike—will learn the tools, best practices, and challenges involved in building a real-world ML application step by step. https://www.oreilly.com/library/view/building-machine-learning/9781492045106/

  • Online: The Business of Data in Maryland with Yet Analytics and Protenus

    In partnership with TEDCO, we are extremely excited to feature two speakers from the successful Maryland-based data-focused companies, Yet Analytics and Protenus. Agenda ------------------------------------------------- 6:00 PM -- Greetings 6:05 PM -- Building a Successful AI Product from Scratch: A True Story - Chris Jeschke 7:00 PM -- xAPI: Facilitating the Documentation and Communication of Learning Experiences through Data - Shelly Blake-Plock Location ------------------------------------------------- Zoom and YouTube Streaming A link will be sent out prior to the event. Please note that Zoom is capped at 100, so if you do not get into the Zoom, you will be able to watch via YouTube. Talks ------------------------------------------------- Building a Successful AI Product from Scratch: A True Story The Protenus Healthcare Compliance Analytics platform is the leading solution used by our nation’s healthcare institutions for privacy monitoring, drug diversion detection and compliance case management. Our platform is a true AI solution - meaning we’ve spent years engineering data, researching signal and developing models for detecting specific breeds of compliance problems. This talk will cover our growing pains in doing so. The first part will cover how we set ourselves up for success to build our product while simultaneously researching signals and improving our models. The second part will go deep into our use of Random Forests as our preferred classification technique and how we strive for explainable AI as a way to build trust with our customers. I’ll cover some of the theoretical basis for random forests as a primer, then go into our process for applying and tuning them for Protenus’ problems. xAPI: Facilitating the Documentation and Communication of Learning Experiences through Data The Experience API (xAPI) is a technical specification for describing learning and training experiences and defining how these descriptions can be exchanged electronically. An open-source data specification, xAPI was designed by the Advanced Distributed Learning Initiative at the US Department of Defense. Data sources include learning management systems, e-Learning mobile applications, serious games, simulations and synthetic training environments, AR and VR, wearables, and both environmental and biometric sensors. In this conversation, we’ll discuss the unique approach of xAPI as a specification based in the world of semantic technology and talk about how xAPI is implemented for data simulation, analytics, and advanced visualization and reporting in the learning and training space. Speakers ------------------------------------------------- Chris Jeschke is an established technical leader and executive in the Baltimore/D.C. metro area, with a strong experiential background building solutions to some of our nation's most difficult intelligence and health problems. Currently, he is the CTO of Protenus, where he leads their product development, operations, R&D and security. Secretly he wishes he was working with quantum languages right now, but haven’t found the time yet. Chris can be reached at [masked] Shelly Blake-Plock is Co-founder, President, and CEO of Yet Analytics, Inc. — a Baltimore-based startup exclusively focused on xAPI. He is a Senior Member of the IEEE where he is an Officer in the Learning Technology Standards Committee and chairs several standards activities including P[masked] on xAPI Implementation, P[masked] on Cybersecurity and xAPI, and P2247.3 on the Evaluation of Adaptive Instructional Systems. He is Principal Investigator on Yet Analytics contracts at the Advanced Distributed Learning Initiative at the US Department of Defense including the Data and Training Analytics Simulated Input Modeler project and the Data Analytics and Visualization Environment for xAPI.