As the effects of climate change are becoming more evident and whose responsibility falls more on man, the scientific community is doing more studies and research to understand the dimension of the problem. And everything seems to indicate that the problem itself is getting out of hand.
A paper published in Nature and that has used Machine Learning to analyze 100,000 studies on climate change explains: “the number of studies since 1990 that observe the impact of climate has increased by two orders of magnitude each year, and that is already putting the manual assessment of the experts to the limit “.
It is a quantity of studies, data and reports that becomes endless, there is simply too much information. Even so-called meta-analyzes, which combine the results of other analyzes to analyze them statistically, can only cover a few hundred reports.
The solution for the moment is turn to artificial intelligence. ScienceAlert comments that there are already AI tools like BERT that seek to understand the text of documents such as studies in order to draw conclusions more quickly. And the truth is that with AI we can analyze large amounts of data to determine trends that humans would take years to detect, but there is a risk of having false positives and conclusions that also raise doubts.
In other words, and as the Nature study concludes, algorithms can help us a lot but we will always need the human mind to be able to interpret conclusions and data. Its basis is easy to understand (certain gases are warming the planet), but all the small implications are becoming increasingly unattainable.
Image | NOAA