Materials Day 2025

Fast-Tracking Sustainability: Accelerating Innovation with Computation and Automation

As climate change and global competition intensify, so does the pressure to innovate faster. In response, researchers are increasingly combining computational tools for rapid decision-making with robotics that interact directly with the physical world. These approaches span from partially automated workflows to fully autonomous, self-driving laboratories. While the promise of accelerated discovery is attracting growing interest and investment, realizing meaningful progress demands a strategic, long-term perspective — one that integrates infrastructure, research, and education in a coordinated way. This will be the central focus of my talk, viewed through the lens of sustainability.

I’ll begin by reviewing select efforts over the past decade to accelerate research, development, and deployment (RD&D), tracing their evolution from proof-of-concept “toy problems” to real-world examples of sustainable-materials innovation achieved on compressed timelines. Next, I’ll highlight three challenging areas where I believe strategic investment could yield especially high returns: predicting the synthesizability of new materials, improving the stability of promising candidates, and scaling up for manufacturing. Along the way, I’ll share some modest successes from our own work, including the discovery of new perovskite-inspired materials and the optimization of existing ones. Finally, I’ll provide examples of education at different stages of the professional life cycle to prepare the workforce for an AI-enabled future.

Bio: Prof. Tonio Buonassisi works at the intersection of machine learning (ML), automation, and materials science to speed up the discovery and deployment of technologies with broad societal impact. His early work in solar energy and technoeconomic analysis supported dozens of companies and earned him the Presidential Early Career Award for Scientists and Engineers (PECASE). From 2018 to 2021, he served as founding director of the Accelerated Materials Development for Manufacturing (AMDM) program in Singapore — a S$24.7M initiative that demonstrated more than a 10× acceleration in materials development by integrating ML, automation, and simulation. Returning to MIT, he led the Accelerated Materials Laboratory for Sustainability and the ADDEPT Center (2023–2025), a U.S. Department of Energy-funded effort that unites academic and industry partners to develop more durable, efficient, and reproducible perovskite-based tandem photovoltaic modules. Starting 2026, he will assume leadership of CSEM’s Sustainability Business Unit, working on full-stack materials-to-products innovation in Neuchâtel, Switzerland.

DAMP Lab: An Academic Cloud Lab for Designing Biology

Bio: Douglas Densmore is the Tegan Family Distinguished Faculty Fellow, a Kern Faculty Fellow, a Hariri Institute for Computing and Computational Science and Engineering Faculty Fellow, and Professor in the Department of Electrical and Computer Engineering at Boston University. His research focuses on the development of tools for the specification, design, assembly, and test of synthetic biological systems. His approaches draw upon his experience with embedded system-level design and electronic design automation (EDA). Extracting concepts and methodologies from these fields, he aims to raise the level of abstraction in synthetic biology by employing standardized biological part-based designs which leverage domain-specific languages, constraint-based genetic circuit composition, visual editing environments, microfluidics, and automated DNA assembly. This leads to a new research area he calls “Hardware, Software, Wetware Co-design”.

Bayesian Optimization: A Brain for Self-Driving Labs

To drive itself, a self-driving lab needs a brain. Often, that brain is Bayesian optimization. Bayesian optimization algorithms design experiments toward the goal of optimizing a noisy black-box objective function that is expensive or time-consuming to evaluate.  They combine a machine-learning-based surrogate for the objective function (often a Gaussian process) with an acquisition function that uses decision theory to quantify the value of a potential experiment. This talk will give an overview of Bayesian optimization and then discuss three frontier areas important to self-driving labs: (1) Moving beyond the black-box assumption in Bayesian optimization to incorporate “grey-box” information such as materials characterization data or intermediate property measurements;  (2) Incorporating human preferences into design goals via human feedback in the form of natural language (“Create a skincare formulation with anti-aging properties and a luxurious feel” or “This formulation is too heavy”) or pairwise comparisons (“I prefer formulation A to formulation B”); (3) Moving beyond Gaussian processes to neural networks that directly choose the next experiment and can incorporate domain knowledge from the internet.

Foundation Models as Autonomous Reaction Optimizers

Our autonomous AI system Coscientist harnesses transformer-based foundation models to bridge the digital-physical gap in scientific discovery. By integrating large language models with experimental automation tools, we’ve created a system that independently designs, plans, and executes complex chemistry experiments, demonstrated through successful palladium-catalyzed cross-coupling optimization. This presentation will showcase our latest advances in autonomous chemical research, from reaction optimization systems for (bio)catalysis and new molecular representations, to the development of automated workflows for extreme scaling of experimental reaction datasets. These technologies promise to dramatically accelerate the development of novel materials and chemical processes, exemplifying how AI can transform laboratory discovery from concept to implementation.

Bio: Gabe leads the Gomes Group at Carnegie Mellon University, where he joined as an assistant professor in the departments of Chemistry and Chemical Engineering in January 2022. He holds affiliations with CMU’s Machine Learning Department, the Wilton E. Scott Institute for Energy Innovation, the Carnegie Mellon Institute for Strategy & Technology, and the CMU BKSQ Lab. Gabe’s the co-founder of 𝑒!”#$, a consultancy firm for scientific evaluations of foundation models. He also serves as an ad-hoc AI for Science advisor to government agencies, non-profit organizations, and companies, including the UK AI Safety Institute and the Office of Science and Technology Policy at the White House (2023–2024). Gabe hails from Brazil, where he received his BSc in Chemistry with Technological Attributes from Federal University of Rio de Janeiro in 2013, with an academic exchange year at the University of Lisbon, Portugal. He earned his PhD in 2018 from Florida State University, under the guidance of Professor Igor Alabugin, where he also was awarded a LASER Fellowship in 2014 and a 2016 IBM PhD Scholarship. At FSU, Gabe’s research was centered on the relationship between molecular structure and reactivity, focusing on the development and applications of stereoelectronic effects. His PhD work earned him the FSU’s Graduate Student Research and Creativity Award, the ACS COMP Chemical Computing Group Excellence Award, and the selection for the CAS SciFinder Future Leaders Program. In 2019, Gabe joined the University of Toronto and Vector Institute for Artificial Intelligence as a Postdoctoral Research Fellow in the Matter Lab, led by Professor Alán Aspuru-Guzik. In 2020, Gabe was awarded the prestigious NSERC Banting Postdoctoral Fellowship with the project “Designing Catalysts with Artificial Intelligence.” Gabe’s vision has been recognized by Chemical & Engineering News in his selection as one of 2022’s Talented 12 in what he calls “transformative digital molecular design”; in 2024, he was selected as a Scialog Fellow for Automating Chemical Laboratories, a program organized by the Research Corporation for Science Advancement. The Gomes Group aims to merge state-of-the-art machine learning, computational chemistry, and automation for reaction discovery and optimization. The group is pioneering the integration of foundation models into chemical sciences and engineering via the development of intelligent agents that can autonomously design, plan, and execute sophisticated experiments on automated and cloud labs. Their cross-disciplinary pursuits extend into fields such as robotics and biomaterials, ultimately shaping conversational interfaces that democratize access to advanced scientific tools. The Gomes group is part of the NSF Centers for Computer-Assisted Synthesis (C-CAS), Chemoenzymatic Synthesis (C- CES), and Accelerated Photocatalysis (CAPs), as well as the Catalysis Innovation Consortium (CIC).

AI Breakthroughs in Materials Science:  How Close Scientists are to Achieving Scientific Superintelligence

Bio: John Gregoire, PhD is the SVP of Physical Science at Lila Sciences. He’s an expert in accelerated materials discovery.  Lila’s focus on scaling accelerated discovery in pursuit of scientific superintelligence lured him from a 15-year career in academia.

Gregoire completed his PhD in physics from Cornell University. As a Caltech research professor in Physics and Materials Science, Gregoire led a team accelerating scientific discovery by automating critical components of research workflows from synthesis and screening to data interpretation and hypothesis generation.

He’s best known for his work exploring and discovering energy-related materials via combinatorial and high throughput experimental methods. Gregoire has held leadership positions for several research consortia, including leadership of  the photoactive materials team in the Liquid Sunlight Alliance (LiSA), a U.S .DOE Energy Innovation Hub.

His passion for incorporating AI in scientific discovery is rooted in mathematical frameworks for knowledge and intelligence, and the focus on materials is rooted in their importance in establishing a sustainable future. He celebrates each discovery while valuing the intelligence gained from making that discovery…

“because that intelligence can be leveraged for a bunch of other problems that we’re moving into, and the echoes of that intelligence will ring for a long time”
– Latitude Media

Machine-Mediated Teamwork

As AI driven systems increasingly become integrated into team contexts, understanding how even subtle machine behaviors influence human interactions and relationships is critical. In this talk, I explore how simple machine actions—like allocation decisions by robots or language suggestions from generative AI—can profoundly affect interpersonal dynamics within teams. Drawing on empirical studies, I illustrate how machine behaviors impact perceptions of fairness, dominance, cooperation, and affiliation among team members. Findings from experiments involving robotic allocation tasks and AI-generated messaging reveal that while these technologies can improve short-term productivity and interaction positivity, they also introduce significant social costs, potentially undermining the relational dynamics crucial for effective teamwork. I argue for a shift in research and design away from anthropomorphic, “teammate” conceptions of machines towards a more nuanced understanding of their role within human relational networks. By recognizing the subtle but impactful ways machines shape human interaction, we can better design technologies that support sustainable, effective, and healthy team environments.

Bio: Malte Jung is an Associate Professor in Information Science at Cornell University and the Nancy H. ’62 and Philip M. ’62 Young Sesquicentennial Faculty Fellow. He directs the Interplay Research Studio that explores the interplay between people and automation.

Jung’s research foregrounds the emotional and relational dimensions of human-AI and human-robot encounters. His work explores how robots become woven into the fabric of everyday life and work. His Interplay Research Studio aims to bring an art studio practice and sensibility to research in information science and robotics. It aims to develop new methods for studying interaction “in the wild” and collaborates across engineering, design, and the social sciences to ask not only how we build AI systems, but how those systems transform the environments, relationships, and values they enter into.

Malte Jung holds a Ph.D. in Mechanical Engineering (with a minor in Psychology) from Stanford University and a diploma in Mechanical Engineering from the Technical University of Munich. Before joining Cornell, he was a postdoctoral researcher at the Center for Work, Technology, and Organization at Stanford. His research has been recognized with multiple best paper awards across HRI, CHI, CSCW, and UbiComp, and an NSF CAREER award.

Open Source Hardware for Self-Driving Labs

Self Driving Labs (SDLs) depend on automation and robotics. However, these technologies have a high barrier to entry in terms of cost and complexity. In this talk, I will describe how DIY and Makerspace inspired approaches are helping us lower that barrier to automation. In particular, I will describe our experiences with the Science Jubilee project: an open-source hardware motion platform with customizable tools driven by computational notebooks. This low-cost setup has been used by various science labs to prototype and build SDL workflows for applications ranging from chemical engineering to synthetic biology. I’ll discuss some of the opportunities and pitfalls of building your own hardware and describe how others have been getting involved.

Bio: Nadya Peek develops unconventional digital fabrication tools, small scale automation, networked controls, and advanced manufacturing systems. Spanning electronics, firmware, software, and mechanics, her research focuses on harnessing the precision of machines for the creativity of individuals. Nadya directs the Machine Agency at the University of Washington where she is an associate professor in Human-Centered Design and Engineering. Machines and systems Nadya has built have been shared widely, including at the White House Office of Science and Technology Policy, the World Economic Forum, TED, and many Maker Faires and outreach events. Her research has been supported by the National Science Foundation, the Alfred P. Sloan Foundation, and the Gordon and Betty Moore Foundation, and her teaching has been recognized with the University of Washington’s Distinguished Teaching Award for Innovation with Technology. She received the MIT Technology Review’s 35 under 35 award in 2020. Nadya is an active member of the global fab lab community, making digital fabrication more accessible with better CAD/CAM tools and developing open source hardware machines and control systems. She is on the board of the Open Source Hardware Association, the editor in chief of the Journal of Open Hardware, half of the design studio James and the Giant Peek, plays drum machines and synths in the band Construction, and got her PhD at MIT in the Center for Bits and Atoms.

Applications of Euclidean Neural Networks to Understand and Design Atomistic Systems

Electrodeposition of Modular Polymer Thin Films for High-Throughput Materials Discovery

Polymer materials are ubiquitous in low- and high-tech technologies, from protective packaging and coatings to advanced electronics, batteries, and drug delivery. The widespread uses and applications of polymers are due to their almost infinite parameter spaces, from side group and backbone chemistry to crosslink density and topology in polymer networks to (micro)phase separated domains in copolymers and polymer blends. These multiscale structural and compositional versatility of polymers make their holistic study, discovery, and development intractable with traditional approaches. Thus, resource- and time-efficient self-driving labs (SDL) will be essential for future polymer discovery, especially when considering the increasing multi-objective optimization demands put on advanced materials that include sustainability and end-of-life considerations in addition to application-specific performance metrics. While automated liquid-based systems have matured over the years, the autonomous discovery of well-defined solid polymeric materials comes with different challenges: uniform film fabrication methods such as spin coating or blade casting are hard to miniaturize, while facile casting methods such as solvent evaporation yield non-homogeneous film topographies and even spatially heterogeneous compositions. To address this challenge, we developed a polymer electrodeposition strategy that enables the efficient and controlled formation of solid crosslinked polymer thin-film materials. To enable autonomous solid polymer material discovery, we combine our novel fabrication paradigm with modular side-group chemistry that can be altered after film deposition together with an SDL capable of electrodeposition and chemical functionalization, as well as compositional, topographic, and functional characterization in a closed loop. We envision this combined system to accelerate direly needed materials discovery such as fluorine-free omniphobic and protective coatings, as well as provide fundamental scientific insights through data-informed development of models for functional polymer networks and materials.

Bio: Jörg G. Werner received his Diploma (M.S.) in Chemistry in 2011 from the Johannes Gutenberg University in Mainz, Germany, and his Ph.D. in 2016 from the Department of Chemistry at Cornell University under the guidance of Prof. Uli Wiesner. After his postdoc in the group of Prof. Dave Weitz at Harvard University, he started his independent research group at Boston University in 2020 as an Assistant Professor in the Department of Mechanical Engineering and the Division of Materials Science and Engineering. His group focuses on spatially controlled synthesis and bottom-up fabrication methods to study the interplay of functional materials with 3D structures across scales, utilizing and studying polymer self-assembly, phase separation, and electrochemistry. He received a DARPA Young Faculty Award in 2023 with a Director’s Fellowship in 2025, an Early Investigator Award in 2025from the PMSE division of ACS, and two Dean’s Catalyst Awards from BU’s College of Engineering in 2020 and 2022. Prof. Werner is a core faculty of the Institute for Global Sustainability and co-founded BU’s Energy and Sustainability Technology (BEST) Lab in 2023, a multi-PI virtual lab that serves as a platform and community for students and faculty to exchange ideas and accelerate their research by open collaboration.