ODSC East 2023 might be over, but you can still find recordings of the virtual sessions on-demand on Ai+ Training. You can watch all of the sessions outlined below here for free, and you can watch the keynotes and demo talks for free here.
Interactive Explainable AI
Meg Kurdziolek, PhD | Staff UX Researcher | Intrinsic.ai
Although current explainable AI techniques have made significant progress toward enabling end-users to understand the why behind a prediction, to effectively build trust with an AI system we need to take the next step and make XAI tools interactive. This session explores how people build trust and the interaction design techniques that can be applied to XAI services.
Reasoning in Natural Language
Dan Roth, PhD | Professor at University of Pennsylvania | VP/Distinguished Scientist at Amazon AWS
This session addresses a few of the challenges of making natural language understanding decisions that are dependent on multiple, interdependent, models. Also discussed is some of the current work in this area including applications, reasoning via decomposition, and navigating information pollution.
Applying Responsible AI with Open-Source Tools
David Talby, PhD | CTO | John Snow Labs
This session explores open-source tools for Responsible AI and some of their real-world applications with a focus on four common challenges: robustness, detection of and fixing labeling errors, bias, and data leakage. You’ll leave this session with an understanding of the available resources for building AI systems that work safely and correctly.
Containers + GPUs In Depth
Emily Curtin | Staff MLOps Engineer | Intuit Mailchimp
This talk is a highly technical walkthrough of the lower-level system libraries involved in GPU computing. By the end you will have covered:
- An overview of the call stack from container, orchestration framework, OS, and all the way down to real GPU hardware
- How ML Eng at Mailchimp provides GPU-compatible dev environments for many different projects and data scientists
- An experienced take on how to balance data scientist’s human needs against heavy system optimization
Winning The Room: Creating And Delivering An Effective Data-Driven Presentation
Bill Franks | Director of the Center for Statistics and Analytical Research | Kennesaw State University
Explore best practices and concrete strategies for creating and delivering a data-driven presentation to a non-technical audience in this talk. You’ll learn how to maximize comprehensibility while maintaining accuracy and audience engagement.
The 10 Ways Machine Learning Systems Can Fail and How to Avoid Them
Bhaktipriya Radharapu | ML Tech Lead, Responsible AI | Google
This ODSC East virtual session looked at different ways in which Machine Learning systems can be compromised, resulting in bias, security concerns, and inaccurate predictions. Some of the ways addressed include concept drift, model inversion attacks, and data poisoning attacks. Lastly, this session explored how to create systems that can withstand these issues.
You can check these sessions out here. And, if you are looking for even more hands-on training sessions, expert-led workshops, and talks, join us at ODSC Europe on June 14th and 15th. Register now to get 40% off any in-person or virtual pass.