Cohousing communities help prevent social isolation

From PBS Groups in Denmark and the U.S. are choosing to live in intentionally intergenerational communities, which emerged to strengthen social ties between aging seniors and their younger counterparts who are balancing work and family. People living in them say the model fosters an interdependent environment and helps everyone feel more comfortable with the process of getting older. NewsHour Weekend's Saskia de Melker reports.

HANDLE robot - wheeled robot from Boston Dynamics

From, Boston Dynamics : Boston Dynamics is best known for its bipedal and quadrupedal robots, but it turns out the company has also been experimenting with some radical new tech: the wheel. The company’s new wheeled, upright robot is named Handle (“because it’s supposed to handle objects”) and looks like a cross between a Segway and the two-legged Atlas bot. Handle robot hasn’t been officially unveiled, but was shown off by company founder Marc Raibert in a presentation to investors. Footage of the presentation was uploaded to YouTube by venture capitalist Steve Jurvetson. Raibert describes Handle as an “experiment in combining wheels with legs, with a very dynamic system that is balancing itself all the time and has a lot of knowledge of how to throw its weight around.” He adds that using wheels is more efficient than legs, although there’s obviously a trade-off in terms of maneuvering over uneven ground. “

Ed Thorp, professor blackjack

From NPR : Ed Thorp was the first 'quant', the first person to make mathematical analysis and statistics the center of his investing. But he only got there because of a card game. As a young man, Ed Thorp was a mathematician doing pretty much what you'd expect a mathematician to do: teaching, studying, trying to solve hard problems. There was one particular problem that nobody else had been able to solve. He wanted to come up with a mathematical system to beat the casino at blackjack.

Google's rules of machine learning - Martin Zinkevich

Rules of machine learning: best practices for machine learning engineering [pdf], by Martin Zinkevich

From : Hard-won lessons, good advice for avoiding mistakes, and rules of thumb. Eg: The number of feature weights you can learn in a linear model is roughly proportional to the amount of data you have.


Subscribe to Ricardo Martins RSS

Scholarly Lite is a free theme, contributed to the Drupal Community by More than Themes.