News, Science & Research

Science & Research Roundup, Sept. 25

By
Science & Research Editor
Monday, September 24, 2018

U. awarded grant to help expedite atom-level simulations

Since atomic-level simulations are becoming more accurate and growing more complex, it takes more time for them to run their courses. Machine learning may be the key to speeding up these simulations by “magnitudes,” said Andrew Peterson, an associate professor in Brown’s School of Engineering, in a University news release.

Research spearheaded by University scientists is funded by a $3.5 million grant from the United States Department of Energy and will take place over the course of four years. The research will initially involve electrochemical reactions and utilize a software called the Atomistic Machine-learning Package, which was developed by Peterson’s research group.

All of these different simulations, including ones that analyze chemical reactions and material properties, run on similar calculations, so machine learning may be able to identify these underlying calculations and figure out how to speed them up.

Researchers examine brain perceptions of optical illusions

University computer experts conducted research to understand the neural operations that go into comprehending a certain type of optical illusion. The team used a computational model, which was subject to the bounds of different parts of the visual cortex, to see how neurons interacted with one another when viewing the illusion.

The study focused on contextual phenomena, a type of optical illusion in which context influences the perception of an image. One example of this illusion is color differentiation, in which one color is contrasted with differently hued backgrounds, which results in differences in color detection. The team created a model of hypothesized feedback connections of neurons, which respond differently depending on the context. The team found that their model’s reaction to the optical illusions mirrored that of humans.

Mothers’ marijuana use tied to children’s earlier use

New research found that the children of mothers that use marijuana between their child’s birth and 12th birthday start using the drug two years earlier than their peers. The study was co-conducted by Natasha Sokol, a postdoctoral fellow in the Center for Alcohol and Addiction Studies, and was based off of longitudinal surveys from across the United States.

Sokol conducted the research as a graduate student, and the study was published in the American Journal of Preventive Medicine Sept. 24. Cassandra Okechukwu, a researcher at Harvard, worked with Sokol on the study.

Topics: