SciTech Roundup Mar. 20
CMU Research Roundup
Using AI to optimize chemistry experiment conditions
As science continues to advance, the number of publications that scientists must sift through grows. Chemists looking to optimize the conditions of an experiment either have a large team to help them go through relevant literature or spend much more time on their own. Hence, the Air Force Research Laboratory (AFRL) has created a challenge aiming to solve this: the Active AI Planners for Chemistry/Materials Optimization and Discovery Grand Challenge.
Researchers at Carnegie Mellon and University of North Carolina at Chapel Hill won Phase I of the challenge in February, pitching a solution that combines expertise in data science and chemical synthesis. But in order to secure the entire $500,000 prize, they must successfully complete the other three stages of the challenge: (2) develop an algorithm, (3) evaluate the algorithm's performance compared to alternatives, (4) develop a program using the algorithm.
The researchers only have the next nine months to complete all phases of the project, so they must move quickly. Hence, the team is working in parallel, 3D printing materials while simultaneously parameterizing the chemical building blocks involved. Then they plan to connect the algorithm to a robot, which will perform experiments autonomously with the goal of surpassing human capabilities.
Ultimately, the project sponsor AFRL hopes that the research can serve as an example of AI-driven optimization that could extend into other fields or problem areas.
Sometimes NASA needs to change the design of spaceflight equipment to suit new requirements or environmental conditions. Imagine an astronaut on the moon that needs equipment parts that are expensive and difficult to prefabricate on Earth. One way of creating such equipment is additive manufacturing, or 3D printing, in which parts are constructed layer by layer based on a 3D model. However, additive manufacturing requires high-quality models, called digital twins, that accurately predict characteristics such as material stress.
To better understand how additive manufacturing can be better utilized in space manufacturing, NASA announced on Mar. 16 that it created two new Space Technology Research Institutes (STRI). One STRI, the Institute for Model-based Qualification & Certification of Additive Manufacturing (IMQCAM), is to be led by Carnegie Mellon and co-led by Johns Hopkins. It is expected to span five years, during which the institute will create digital twins of spaceflight materials using physics, mechanics, and machine learning. Once the digital twins are evaluated against experimental material data, they can then be used in digital simulations that will test which material performs better in fractions of the time it would take to do in the real world.
NASA has also created another STRI, Quantum Pathways Institute, led by the University of Texas at Austin. This institute will focus on advanced quantum sensing technology, which uses quantum science principles to get better measurements of data such as the Earth's change in mass.
Each STRI has been allocated $15 million, funded by Space Technology Mission Directorate's Space Technology Research Grants. Two STRIs have been created every other year, each with the mission of strengthening NASA's ties to the academic community.
Rales Foundation and CMU invest $150 million into new scholarship program
Carnegie Mellon and the Norman and Ruth Rales Foundation announced the creation of the CMU Rales Fellows Program, a scholarship aimed at addressing the "Missing Millions" — what the National Science Foundation hypothesizes to be millions of people who cannot contribute to the STEM field due to racial, ethnic, gender, and socioeconomic barriers. The CMU Rales Fellows Program will provide full tuition as well as holistic support such as professional networks to an estimated 80 fellows in M.S. and Ph.D. STEM programs annually, hoping to develop "a diverse community of exceptionally talented national STEM leaders from underrepresented or under-resourced backgrounds … including first generation students and students from low socioeconomic backgrounds."
The Rales Foundation has endowed $110 million to support the program, Carnegie Mellon has invested $30 million, and both have established an additional $10 million fund. To help with recruiting fellows and advise potential fellows, Carnegie Mellon will be partnering with the Ron Brown Scholar Program.