Categories

Knowledge

Can cars in the streetscape not only distinguish the gap between the rich and the poor but also the political orientation?
Jan 17, 2019

 Recently, the Visual Studio of Stanford University published a paper on artificial intelligence applications to analyze the political orientation of the residents and the gap between the rich and the poor by analyzing the car information in Google Street View . It is understood that if the number of cars in a community is much larger than the number of pickups, then the region has an 88% chance of voting for the Democratic Party; conversely, if the number of pickups is greater than the number of cars, the vote for the Republican Party may be 82%. What is going on?

  Artificial intelligence improves research efficiency

  According to Huaqiang Wisdom, the paper was published by researchers who selected the 200 most densely populated cities in the United States to identify and judge vehicles appearing in 50 million Google Street Views by establishing an artificial intelligence algorithm- computer vision technology system . After the information, the conclusion is drawn.



  In order to improve the accuracy of the algorithm, the researchers obtained detailed photos of 15,000 cars from the car sales website and established a database of all models sold since 1990. The next step is to classify the car information in Street View by brand, model and year. This number is painful for researchers because when you collect and mark small differences in the vehicle, sometimes you don't notice the difference. It took only two weeks for artificial intelligence to complete the classification of 220,000 cars. It would take at least 15 years for a person to do it.

  AI predicts political inclination

  After collating the data, the researchers compared the current most comprehensive population database, US community surveys, and presidential vote data, and found a simple linear relationship between auto, demographics, and political tendencies.

  According to systematic research, if the number of cars in a community is much larger than the number of pickups, then the region has an 88% chance of voting for the Democratic Party. If the number of pickups is greater than the number of cars, the probability that the constituency will vote for the Republican Party is 82%.

  In addition, the study found that the most environmentally friendly city is Vermont, and Chicago is the most disparate city between the rich and the poor, with expensive sports cars and cheap cars crowded on the streets. New York has become the city with the highest per capita vehicle price.

  Understanding society to improve life



  According to the researchers, this computer vision technology system can help us understand how society works, what people need, and how to improve their lives. For example, monitoring carbon dioxide levels and mitigating traffic congestion; providing a more timely and continuous supplement to current demographic surveys. Be aware that human-based home surveys cost 2.5 billion a day and the results are delayed.

  The potential of this technology will be enormous. In September 2017, Google made the biggest upgrade to its own streetscape. While improving the image quality, it also began to use artificial intelligence technology to enhance the identification of information such as roadside store abbreviations and capitalized logo names. The future will serve the society better.

  Huaqiang Wisdom Network Conclusion

  Data is the raw material for artificial intelligence, and Google is the rich player. With the application of artificial intelligence technology, more and more simple connections that we have not noticed are discovered and applied to social services. But all this depends on the analysis of massive data, and the future may indeed predict political trends.

  • QR CODE
  • facebook