Over the past couple of years, the consumer robotic market has seen rapid adoption due to the pandemic and dropping prices, and the market is expected to grow at a CAGR of almost 31% through 2027. Last year, Amazon announced Astro, a new household robot for home monitoring that combined AI, computer vision, sensor technology, and voice and edge computing. Astro brought new advances to the consumer robotic market around human-robot interaction and multimodal AI. In this fireside chat, Dr. Ken Washington, VP of Software Engineering for Consumer Robotics, will discuss scaling robotics at Amazon and new features in Astro that are novel in the areas of computer vision, perception training, and mapping. Dr. Washington is joined by Vijay Karunamurthy, Head of Engineering at Scale AI, to discuss the future of robotics at Amazon and the new AI technologies evolving with consumer robotics. Before joining Amazon, Dr. Washington was CTO at Ford Motor Company, where his portfolio included propulsion systems, sustainable and advanced materials, additive manufacturing, next-gen vehicle architectures, controls, and automated systems. Dr. Washington also served as Lockheed Martin Corporation’s first chief privacy officer and as CIO for Sandia National Laboratories. In addition to his election to the NAE in 2020, he received the 2012 Black Engineer of the Year Award in Research Leadership.

James Manyika, Senior Vice President for Technology and Society at Google, will sit down with Alexandr Wang, Scale founder and CEO, to discuss the best ways to ensure that AI models assist work and grow the economy, even as they replace certain mundane tasks. Manyika will describe how prioritizing utility and focusing on harm reduction should be at the core of principled use and deployment of AI. Manyika will point out the need to allow people to know whether they are interacting with an AI agent or a human, versus scenarios in which it’s permissible to use algorithms and human workers interchangeably. Learn how to adapt to new forms of work and activities that evolve as AI systems become more capable, how to develop AI policy to ensure no harm comes from the use of AI, and about the geopolitical impact and importance of AI principles. Manyika will also discuss his recent work for the Academy of Arts and Sciences, as editor of the Spring 2022 issue of Dædalus. For this journal, he invited a number of noted AI researchers and tech luminaries to write about different aspects of AI and society, including how the technology is being used for climate change, healthcare, and the economy—particularly how it can be made even better in those use cases.

Admiral William H. McRaven, a retired Navy four-star admiral and former SEAL, sits down with Scale Founder and CEO Alexandr Wang to discuss why America must regain its dominance in technology innovation – and how, by acting decisively, it can do so. In recent years, some other countries have outspent the U.S., but an effective policy is about much more than budgetary considerations. Admiral McRaven discusses leadership lessons that apply to technology strategy based upon his decades of experience in special operations, his combat during Desert Storm and the Iraq and Afghanistan wars, his organization and leadership of the team that captured Osama bin Laden, and his years of advising U.S. presidents George W. Bush and Barack Obama on defense issues. The Admiral has also written several books that wound up on The New York Times’ best-seller list, including Sea Stories: My Life in Special Operations; The Hero Code: Lessons Learned from Lives Well Lived; and Make Your Bed: Little Things That Can Change Your Life…And Maybe the World. The last was originally a commencement speech that the admiral repurposed into a book. Admiral McRaven is on the boards of the Council on Foreign Relations (CFR), the National Football Foundation, the International Crisis Group, The Mission Continues, and ConocoPhillips.

Jeff Wilke is chairman and co-founder of Re-Build Manufacturing, a private company helping to bolster America’s industrial competitiveness. Before taking on this role, Wilke spent more than 21 years at Amazon, most recently as CEO of Worldwide Consumer, where he grew the business from $1 billion to a mind-boggling $386 billion a year. He spent his first seven years at Amazon building global operations. Wilke brings a wealth of knowledge across areas including supply chain, retail, and operations. Wilke speaks with Scale CEO Alexandr Wang to discuss his journey at Amazon and covers how focusing on data and building models can help power eCommerce marketplaces. Learn how Wilke applied AI across the entire eCommerce stack and created an authoritative and comprehensive product catalog with more than 12 million listings. He will also share his predictions about the future of AI and sustainability. Wilke is an active philanthropist and mentor of startup founders who identify as female and/or People of Color. He is Vice Chairman of Code.org and Chairman of the U.S. Government’s Advisory Committee on Supply Chain Competitiveness.

Stable Diffusion disrupted the deep learning scene when it was released in August, advancing the field of text-to-image models because of its ability to generate photo-realistic images given any text input. And unlike other models, Stable Diffusion makes its source code available, further advancing the democratization of AI. Stable Diffusion was created by Stability AI, in collaboration with EleutherAI and LAION. Stability AI is on a mission to design and implement solutions using collective intelligence and augmented technology and has developer communities with over 20,000 members who are building AI for the future. In this fireside chat, Emad Mostaque, CEO of Stability AI, will discuss emerging trends for open source AI infrastructure, the importance of data for real-world applications of AI, and predictions on the development of “text-to-everything” in artificial intelligence.

Content comes in many forms—from traditional media with professionally developed content to social media with user-generated content, and everything in between. Over the past decade, media platforms have evolved significantly to keep up with the pace of newly developing content and content trends. For example, social media now support multiple forms of content, including images and short-form video, and traditional media platforms have expanded to support in-house content and streamed content. This expert panel will explore the importance of recommendation systems, discuss the importance of the data that powers these systems in the ever-evolving media platform experience, and cover how and why recommendation systems are so critical for growing media-based businesses.

Join this enterprise-focused, spirited discussion on how best to train, use, and fine-tune foundation models in the enterprise. Elliot Branson, Director of Machine Learning & Engineering, Scale AI, will moderate the panel with industry experts from AWS, NVIDIA, Netflix, and Meta. Erhan Bas, formerly Applied Scientist at Amazon Web Services and now at Scale, shares his perspective on training large language models (LLMs). Bryan Catanzaro, Vice President of Applied Deep Learning Research at NVIDIA, shares how the GPU manufacturer is targeting foundation models as a core workflow for enterprise customers. Faisal Siddiqi, Director of Machine Learning Platform at Netflix, will share how his company is using foundation models to analyze highly produced video content. Susan Zhang, Researcher at Facebook AI Research (FAIR), a division of Meta, will share insights from training and fine-tuning Meta’s OPT model. Members of the panel will share how they scale their training across multiple nodes, attempt to avoid overfitting by mitigating data quality issues early on, and address bias in models trained on a large internet-based text corpus. The panelists will discuss the compute cost inherent in training an LLM from scratch, how to avoid costly and tedious hyperparameter optimization, the need to mitigate training failure risk in clusters with thousands of GPUs, including sticking to synchronous gradient descent, and the need for extremely fast storage devices to save and load training checkpoints.

Machine learning leaders from robotics (Covariant), home automation (Resideo), autonomous delivery (Nuro), and warehouse automation (Pickle Robot) sit down with Russell Kaplan, Scale’s Director of Engineering, to share their approaches to dataset management. Pickle Robot CTO Ariana Eisenstein will share how she thinks about modulating quantities from different data sources like synthetic and public open datasets with real-world data for training datasets. Mostafa Rohaninejad, Founding Research Scientist at Covariant, will describe how the object “picking” problem requires synthetic data for unsafe scenarios and how he also incorporates structured and time-series data—supervised and unsupervised learning should go hand-in-hand. Jack Guo, Head of Perception at Nuro, will explain how it’s essential to have tools and mechanisms to automatically highlight recorded data that deviates from the norm, especially if it was captured in a new location. Like Rohaninejad, he will stress the importance of simulation as a component of successful reinforcement learning. Louis Tremblay, AI/ML Engineering Leader at Resideo, will explain how security cameras in the home represent an even more unbounded environment than do warehouses. The group will also discuss why maintaining separate datasets and training pipelines for different customers is both costly and incurs additional technical debt over time. Testing on fault-tolerant customers first before deploying to the wider fleet is also important. Scale’s Kaplan will share how, in his experience, when metrics and anecdotes seem at odds, it makes sense to re-think the metrics and establish new ones.