BIDMaP is happy to be jointly hosting this year's seminar series with colleagues in computational physical sciences, including collaborators from the departments of Physics and Astronomy, and the Lawrence Berkeley National Laboratory.
Some of the most powerful techniques developed in ML are rooted in physics, such as MCMC, Belief Propagation, and Diffusion based Generative AI. We have recently witnessed that the flow of information has also reversed, with new tools developed in the ML community impacting physics...
Atomic systems (molecules, crystals, proteins, etc.) are naturally represented by a set of coordinates in 3D space labeled by atom type. This poses a challenge for machine learning due to the sensitivity of coordinates to 3D rotations, translations, and inversions (the symmetries of 3D Euclidean...
Density functional theory (DFT) lies at the heart of all practical applications of theoretical chemistry. However, it is well-known that DFT often provides unsatisfactory descriptions of many important systems: non-equilibrium geometries, such as transition states, and strongly correlated systems, such as transition metals and excited...
Please join us for an interdisciplinary discussion of scientific themes addressed in this series. We'll have lightning talks by Saumil and Théo (BIDMaP Fellows) to start, and then an informal meeting in the seminar space with tables for small or large group discussions.
Propylene is an important building block for the manufacturing of various chemicals and plastic products. The ever-increasing propylene demand is hardly met by traditional oil-based cracking processes, known for their high energy consumption and substantial CO 2 emissions. Leveraging the abundance of light alkanes from...
The pursuit of carbon neutrality has become a global imperative in the face of climate change, driving the transition to renewable energy sources and the widespread adoption of electric vehicles. Designing new cathode materials for energy storage is one promising avenue. Modern battery materials such...
Perovskite is one of the most promising photovoltaic materials for the future. While low stability has long been the bottleneck issue limiting their commercialization. In the past, by using enhanced sampling coupled with machine learning (ML) potential model, I have unraveled the degradation mechanism of...
The advent of advanced large language models like ChatGPT marks a transformative era in scientific research, particularly in the field of reticular chemistry. This seminar focuses on how ChatGPT's natural language processing capabilities enable scientists to accelerate and innovate in their research endeavors. We will...
Seismic networks have consistently improved across extensive temporal and spatial scales, enhancing our capability to record subtle shaking signals of the Earth. The vast seismic archive poses a challenge for efficient data analysis, but also presents an opportunity to uncover many hidden signals from small...
Abby Doyle is a professor of chemical and biomolecular engineering at the University of California, Los Angeles (UCLA). Her research is focused on tackling unresolved issues in the field of organic synthesis by creating innovative catalysts, catalytic reactions, and synthetic techniques. Recently, the Doyle group...
Bio Alán Aspuru-Guzik is a professor of Chemistry and Computer Science at the University of Toronto and is also the Canada 150 Research Chair in Theoretical Chemistry and a Canada CIFAR AI Chair at the Vector Institute. He is a CIFAR Lebovic Fellow in the...
In silico materials design often involves the exploration of vast, diverse chemical spaces. While ab initio methods have been phenomenally successful in materials simulations, their scope of application has always been constrained by their high cost and poor scaling. In this talk, I will highlight...
The success of deep learning has hinged on learned functions dramatically outperforming hand-designed functions for many tasks. However, we still train models using hand designed optimizers acting on hand designed loss functions. I will argue that these hand designed components are typically mismatched to the...
Please join for this EECS Colloquium (not part of the regular BIDMaP series; please note different day/time/location). Many thanks to EECS for welcoming the BIDMaP community to this exciting talk. See more on the EECS website, including a remote participation option.
I will describe a research program aimed at advancing the potential for discovery and interdisciplinary collaboration by approaching fundamental physics challenges through the lens of modern machine learning (ML). This research program has two complementary components. Ab initio simulations are a powerful tool of fundamental...
