Keynotes

Modern Hybrids

Abstract

This talk summarizes our 15+ years of work on the use of Machine Learning for Search & Optimization. I review the four main approaches that we invented during this time. Since learning during search takes effort, it should not surprise that we designed three of these approaches for a particular target range of total function evaluations: from few tens of dozens, to thousands, to many hundreds of thousands of function evaluations. The last hybrid I review regards a surrogate-based approach for optimization under stochastic uncertainty. The wonder of this research area is that each of these four methods defines the state of the art in its respective area, giving significant empirical evidence that learning to optimize can be highly effective.

Speaker Bio

Meinolf currently serves as CTO of InsideOpt, a US-based startup that produces general-purpose software for automating decision-making under uncertainty. Before, he held positions as Director for Network Optimization at Shopify, Lab Director for the Machine Learning and the Knowledge Representation Labs at General Electric's Global Research Center, Senior Manager for Cognitive Computing at IBM Research, and Assistant Professor for Computer Science at Brown University. Meinolf received his doctorate degree in 2002 from Paderborn University (Germany) and then went on to Cornell University as Postdoctoral Associate. Meinolf has published over 80 articles in international conferences and journals, holds six patents, served as PC Chair of IAAI 2021 and 2022, LION 2016, and CPAIOR 2013, Conference Chair of CP 2007, and Associate Editor of the Informs Journal on Computing. He won over 22 first prizes at international programming competitions, most recently two first prizes at the AI4TSP competition held at IJCAI 2021.

Emotion, Social Robots, and a New Human-Robot Relationship

Abstract

People have welcomed conversational AI technologies into our homes, workplaces, and institutions where we interact with them on a daily basis. The proliferation of digital assistants in a multitude of embodiments (e.g., speakers, displays, avatars, robots) in human environments over extended periods of time provides us with new ways to investigate, develop and assess the design of personified AIs that emotionally engage and support people to promote human flourishing across a wide range of applications and usage contexts. In this talk, I highlight a number of research projects where we are developing, fielding, and assessing social robots in homes, schools, and living communities of older adults. We explore different embodiments and develop adaptive algorithmic capabilities for our robots to sustain interpersonal engagement and personalize to people’s needs to support novel interventions in education, social engagement, and emotional wellness. In addition to evaluating the impact of these capabilities and features on improving learning, sustaining engagement, nudging behavior, and shifting attitudes — we are also examining the nature of the relationship that people form with these personified AI technologies and how it contributes to these impacts. We conclude by reflecting on the ethical and responsible design of intelligent technologies that emotionally engage and build relationships with people.

Speaker Bio

Cynthia Breazeal is a Professor at the MIT Media Lab where she founded and directs the Personal Robots Group. She is also MIT dean for digital learning leading professional education, and director of MIT’s initiative on Responsible AI for Social Empowerment and Education (RAISE) to help bring AI education to K12 and the workforce. She is a pioneer in the field of social robotics and human-robot interaction. Her research focuses on the design and real-world impact of personalized and emotionally engaging personified AI technologies that promote personal growth, learning, creativity and flourishing by people of all ages. She is author of the seminal book ‘Designing Sociable Robots,’ named a AAAI Fellow, and is a recipient of the George R. Stibitz Computer & Communications Pioneer Award. She has spoken at prestigious venues such as TED, CES, SXSW, the World Economic Forum, and the United Nations on topics related to AI, innovation, and society. She is globally recognized as an award-winning innovator, designer, and entrepreneur. Her work has been recognized by the National Academy of Engineering, the National Design Awards, and Technology Review’s TR100/35 Award. She was founder, Chief Scientist and Chief Experience Officer of the mass consumer home robotics startup, Jibo, Inc. whose eponymous robot received numerous design and innovation awards by CES, Fast Company, Core 77, and was featured on the cover of TIME magazine as part of the 2017 Best Inventions Awards. She received her doctorate from MIT in Electrical Engineering and Computer Science in 2000.

An Evolutionary Optimizer’s Path to Commercial Success and Some Rocket Science Beyond It

Abstract

Few EC technologies have gone from universities to commercial success. Goodman will describe the SHERPA algorithm, part of the HEEDS design exploration framework, and how Red Cedar Technology, which he co-founded, eventually succeeded. Beginning 20 years ago, SHERPA used a self-adapting ensemble of EC methods (GA, ES, DE, SA, etc.) in each run, requiring no choice of optimization methods or parameters by the engineering user. It is a best-selling engineering design optimizer, still built around the original code developed in 1999-2010, although current owner Siemens now has 20+ developers on HEEDS and SHERPA. Goodman will then turn to a problem outside SHERPA’s scope, addressed with unpublished parallel EC methods. NASA provided a futuristic challenge problem to teams of DARPA awardees, to develop ways to optimize the distribution of a set of solid propellant types (eventually to be 3D-printed) in a rocket. Goodman will describe the modeling of the rocket and several problem-specific EC methods used to find feasible solutions to this problem with a design space of 10^(300-500) and over 700 constraints.

Speaker Bio

Erik D. Goodman is PI and Executive Director of the BEACON Center for the Study of Evolution in Action, an NSF Science and Technology Center headquartered at Michigan State University, funded by NSF for 2010-20, and now continuing with funding from MSU. BEACON has a dynamic research program and extensive education and outreach programs, and includes evolutionary biologists as well as computer scientists/engineers studying evolutionary computation (for search and optimization) and evolution of digital organisms. Goodman is a professor in Electrical and Computer Engineering, Mechanical Engineering, and Computer Science and Engineering. He was co-founder and VP Technology, Red Cedar Technology, Inc., (now a division of Siemens), which developed design optimization software that has become a best-selling system in industry. He was named Michigan Distinguished Professor of the Year, 2009, and received the MSU Distinguished Faculty Award in 2011. He was elected Chair of the Executive Board (2003-2005) and Senior Fellow, International Society for Genetic and Evolutionary Computation; then was Founding Chair of the ACM SIG on Genetic and Evolutionary Computation (SIGEVO), 2005. His current personal research is on evolutionary algorithms for optimization of heterogeneous propellant grains for solid-fuel rockets and on evolutionary approaches to neural architecture search.