NVIDIA GTC is the must-attend digital event for developers, researchers, engineers, and innovators looking to enhance their skills, exchange ideas, and gain a deeper understanding of how AI will transform their work. Registration is available here https://reg.rainfocus.com/flow/nvidia/gtcfall20/reg/login
Modzy Presenting at NVIDIA GPU Technology Conference
BETHESDA, Md.–(BUSINESS WIRE)–Modzy, the leading enterprise AI platform, today announced that it will be a Global Diamond sponsor of the upcoming NVIDIA GPU Technology Conference (GTC), with workshops and the conference scheduled for Oct. 5 – 9. NVIDIA GTC is the must-attend digital event for developers, researchers, engineers, and innovators looking to enhance their skills, exchange ideas, and gain a deeper understanding of how AI will transform their work. Registration is available here.
“Modzy was introduced at NVIDIA GTC DC last November,” said Josh Sullivan, head of Modzy. “I’m humbled by the momentum we’ve achieved in less than a year in technology innovation and building our business. We’re proud that our people, our greatest asset, are featured in GTC programming this year.”
According to a Forrester report written by Kjell Carlsson and Mike Gualtieri, “Given the rapid developments in ModelOps offerings, many enterprises will be tempted to wait until the offerings have matured and the market has stabilized. That is not an option. To unlock the value of their AI initiatives today, all organizations need to rapidly improve their ModelOps capabilities using the options that are available to them now.”1
Introduced by Booz Allen Hamilton in 2019, Modzy is the enterprise AI platform to easily secure, govern, deploy and manage trusted AI. Modzy’s participation at NVIDIA GTC includes:
Featured Session
Three Reasons Your AI Will Fail: Model Drift, Adversarial Attacks, and Lack of Explainability
Tuesday, Oct. 6, 9:00 – 9:50 EDT
Dr. Josh Sullivan, Head of Modzy
Dr. Arash Rahnama, Head of Applied AI Research, Modzy
It’s no secret the AI you’re running in production today is brittle, susceptible to adversarial attacks, and needs to be more transparent and effective. Couple that with sophisticated attackers leveraging techniques such as data poisoning to twist model outputs, and your models are completely vulnerable to both inside and outside threats. What steps can you take now? We’ll show you how to leverage cutting-edge techniques to achieve better performance, powered by GPUs. Expect to walk away with the latest techniques you can implement now to defend against attacks and achieve a new level of performance with your AI specifically related to:
Monitoring drift detection
Embedding adversarial defense, and
Increasing explainability
On-Demand Sessions
Closing the Gap in Real-world AI: Automating Model Deployment with ModelOps
Dr. Gabriella Melki, Lead Data Scientist, Modzy
Saumil Dave, Lead Data Scientist, Modzy
Learn how to deploy, manage, monitor, retrain, and secure models in production faster than ever before using ModelOps and (1) automated model deployment, and (2) model retraining, The hardest part of AI shouldn’t be the hand-off between data science and development teams, and this talk will explore ways to reduce friction in this process, and speed up and improve how you build AI-powered applications. Examples will show an automated deployment pipeline, powered by the NVIDIA CUDA platform, and model retraining using a Jupyter notebook powered by an NVIDIA V100 GPU.
Explain This: Beyond Lime and SHAP, the Fastest Approach to AI Explainability
Andrew Tseng, Research Data Scientist, Modzy
Learn how a novel approach to explainability based on adversarial machine learning can be used to explain the predictions of deep neural networks, powered by a single GPU Tesla V 100-SXM2 to produce better results faster than LIME and SHAP. We’ll cover our approach, which identifies the relative importance of input features in relation to the predictions based on the behavior of an adversarial attack on the DNN and uses this information to produce the explanations. We’ll include examples that demonstrate how explainability accompanies AI security, and compare the speed and performance of this new approach to other leading explainability techniques.
Making AI Work for the Public Sector and Highly Regulated Industries
Josh Elliot, Head of Operations, Modzy
Join industry leaders who’ll discuss what it takes to effectively deploy, manage, monitor, and secure AI in government agencies and other highly regulated industries. We’ll discuss the tools you need to ensure that your teams are empowered to build and deploy trustworthy AI, all while providing you governance and audit capabilities to understand how and where AI is being used across your enterprise. You’ll walk away with real-world knowledge to manage the risks associated with “shadow AI” and understand the solutions your teams need to build AI-powered applications that will enable time and cost savings, powered by GPUs. We’ll include examples of how you can have insight into AI usage and performance across your enterprise, and how your teams could be deploying models into production in a matter of minutes, rather than months.
About Modzy
Modzy is the leading enterprise platform to secure, govern, and deploy Artificial Intelligence (AI) models. Modzy helps government and private-sector customers meet the challenge of operationalizing AI by enabling rapid deployment, management and governance of trusted AI. Modzy’s platform leverages embedded security, patented adversarial defense, explainability of model predictions, and governance features to help customers easily manage and quickly benefit from their AI investments at enterprise scale. Modzy offers choice with ready-to-deploy pre-trained and trainable AI models from leading companies and open source communities. Modzy accelerates the deployment of trustworthy AI while increasing transparency, lowering the barriers to adopting and scaling AI. To learn more visit www.Modzy.com.
1 Forrester, “Introducing ModelOps To Operationalize AI,” Kjell Carlsson, PhD, Mike Gualtieri, et al, Aug. 13, 2020.