微軟推出新機器學習工具 提升天氣預報準確性
微軟發布開源軟體，利用機器學習提高天氣預報準確性，適用於資源有限的氣象機構。
microsoft has released an open-source machine learning tool to improve weather forecasting accuracy two to six weeks in advance. the tool, called probabilistic bias correction (pbc), corrects systematic bias in existing weather models, making ai-enhanced forecasts more affordable for resource-strapped weather agencies. pbc improved the accuracy of a leading physics-based weather model for 98 percent of precipitation and 92 percent of temperature targets, and doubled the accuracy of the european centre for medium-range weather forecasts' new ai forecasting system. the tool can run on cpus, making it more resource-efficient than traditional physics-based modeling or ai-based models.