Race to the moon

Ever since the Soviet Union crash-landed a probe on the lunar surface in 1959, only the national space programs of the United States, the Soviet Union, and China have had the financial and technological resources to safely land manned or unmanned spacecraft on the surface of the moon—until now. In January, the Google Lunar X Prize competition selected five finalist teams from private companies that will attempt to go to the moon before this year is out.

The teams, representing Israel, India, Japan, the United States, and other countries (plus one international team), must receive at least 90 percent of their funding from private sources. The winning team—with a grand prize of $20 million—must successfully land its spacecraft on the moon, and its lander must travel at least 500 meters (1,640 feet) and transmit high-definition video of its surroundings back to Earth.

The Israeli team and one from the United States plan to use “hopper” landers that will fly, rather than drive, the required 500 meters. Israel’s SpaceIL team was the first to secure a launch contract (aboard a SpaceX rocket). American contestant Moon Express was the first private company in history to receive governmental permission to travel to the moon.

“Our goal,” said Moon Express CEO Bob Richards in a statement, “is to expand Earth’s social and economic sphere to the moon, our largely unexplored eighth continent.”

Say again?

If you didn’t get an Amazon Echo for Christmas, you may know someone who did. The popular device with its voice-operated assistant, Alexa, was a top seller over the holidays, and Amazon has sold more than 5 million since its launch two years ago.

The massive amount of voice data Alexa is collecting from those millions of users is helping Amazon solve a long-standing speech recognition problem known as the cocktail party challenge: picking out a single voice in a roomful of many people talking.

Developers at Amazon created a version of Alexa that notes distinguishing characteristics of a voice calling its name and then homes in on the words of that person, according to MIT Technology Review. They hope eventually to develop a version that can follow two people speaking simultaneously. —M.C.

Sounding stressed

Disorders affecting mental health, such as post-traumatic stress disorder (PTSD) and traumatic brain injury (TBI), are difficult to diagnose. But researchers believe certain vocal cues extracted from voice samples could be used to detect the presence of these and other diseases.

Charles Marmar, chairman of the department of psychiatry at NYU School of Medicine, is conducting a five-year study using artificial intelligence algorithms to analyze thousands of voice samples from veterans. His goal is to learn the relationship between the conditions of PTSD, TBI, and depression and vocal markers such as tone, pitch, rhythm, rate, and volume.

In a 2015 study of 39 male volunteers, Marmar and his team developed a voice test that distinguished between PTSD patients and healthy volunteers with 77 percent accuracy, according to MIT Technology Review. Additional voice samples collected since then are helping identify speech patterns that could possibly distinguish between PTSD and TBI.

Researchers have also discovered links between vocal markers and the risk of coronary heart disease. They envision a vocal test on a smartphone app that could function as an inexpensive predictive screening tool for multiple diseases. Responding to concerns about privacy, researchers point out that the algorithms simply capture vocal patterns, without logging what is said. —M.C.