STARS AND GRIPES ... Many in the restaurant industry were in a celebratory mood on Tuesday when the 2021 California Michelin Guide was released, even those who didn't make the list. Selby's in Atherton was among the new additions to the list, earning a respectable one star. Bacchus Management Group opened the swanky, upscale restaurant in July 2019. "Congratulations to our amazing team for working extremely hard under the craziest conditions," the restaurant said in a Instagram post. Other local restaurants held onto their single stars, including The Village Pub in Woodside, Madera in Menlo Park, Protege in Palo Alto and Chez TJ in Mountain View. The jubilance also carried over to Baume in Palo Alto, though not in the way most people would expect. Bruno Chemel, chef and owner of the French restaurant on California Avenue, was grateful to not make this year's guide, which was a goal he's been working toward for several years. "The erratic, corporate focus and fixations of the Guide have been a distraction for years, creating no value for restaurateurs or diners," he said in a press release. "I finally feel free to enjoy cooking just for the happiness of my guests. Instead of following the status quo, I can follow my own vision and listen to the customer's feedback, not the Michelin guidelines."
IMPACTFUL ASTEROIDS ... Palo Alto High School senior Franklin Wang's discovery of six previously undetected near-Earth asteroids recently won him a $50,000 scholarship from the Davidson Institute. Wang created a machine learning model that identifies the asteroids from telescope images. The Davidson Fellows Scholarship Program honors young people who have "completed significant projects that have the potential to benefit society in the fields of science, technology, engineering, mathematics, literature and music," according to a press release. In the past, Wang said he had used machine learning for more light-hearted projects, like creating a song lyric generator. "I wanted to work on a more serious machine learning project that could actually ... improve upon previous works," Wang said. He got the idea for the project after reading a paper from the Zwicky Transient Facility that used machine learning to find asteroids. Photographs of asteroids capture a streak similar to a shooting star, but these streaks can be hard to detect with conventional techniques, Wang said. Machine learning offered a potential path, but the challenge was how "data hungry" these algorithms are, Wang said. "We don't have a very large data set of these asteroid streaks that are available on hand, and because of that it's quite tricky to train these machine learning models," Wang said. He simulated roughly 500,000 asteroid trails using mathematical expressions that he then used to train his model to identify what is, and is not, an asteroid.