Neural network predicts neighborhood voting patterns based on Google Street View images
Researches from Stanford analyzed 50 million photos of street scenes and location data. Their goal was to find out how demographic statistics at the zip code and precinct (approximatly 1,000 people) levels can be predicted.
The analysts gleaned information on what cars appeared in different areas. They were able to detect 22 million cars — 8% of all cars in the country — in 3,000 zip codes and 39,000 voting districts. Hundreds of people from places like Mechanical Turk, as well as car experts trained algorithm to identify vehicles in a sample of millions of pictures.
The researchers made accurate predictions about a neighborhood’s income, race, education and voting patterns based on data about these cars. To do so they cross-referenced the data with information from other sources, including the Census Bureau’s American Community Survey and presidential election voting records.
#interesting For example, extended-cab pickup trucks are most strongly associated with Republican voting districts while sedans — with Democratic districts.
The team wrote that their technology can contribute to the American Community Survey which takes only small part of population because of high costs and labor-intensivity.
Microsoft created artificial intelligence with imagination
The company's new computer artist draws images from a blank page based on text descriptions. As the sentence is read, the algorithm adds new elements of the image.
Important is that AI adds extra details into the picture — something besides the description. This shows the imagination of the algorithm. From the Microsoft announcement:
"If you’re handed a note that asks you to draw a picture of a bird with a yellow body, black wings and a short beak, chances are you’ll start with a rough outline of a bird, then glance back at the note, see the yellow part and reach for a yellow pen to fill in the body, read the note again and reach for a black pen to draw the wings and, after a final check, shorten the beak and define it with a reflective glint.
Then, for good measure, you might sketch a tree branch where the bird rests. Now, there’s a bot that can do that, too."
#interesting The algorithm is able to draw not only regular objects but also absurd ones – like “a floating double-decker bus.”
Nvidia driver assistant for Volkswagen
Volkswagen revealed new feature of its new I.D. Buzz, all-electric retro-inspired camper van concept.
Into I.D. Buzz, as well as into company's other future vehicles, will be built artificial intelligence technology by Nvidia — Nvidia Drive IX Platform. It can enable features like face recognition-based door unlocking, gesture input for cockpit controls, natural language speech recognition and even monitoring a driver’s attentive and distracted state for providing safety alerts to bring them back to focus.
#background Nvidia has been working with other automakers on putting artificial intelligence features, including Audi and Mercedes-Benz.
Also published on Medium.