Machine learning technologies power computers to perform analytical tasks at higher rates of speed than humans. When algorithms and data are fed into computer models, machine learning provides what Forbes contributor H.O. Maycott calls, “an intelligence layer to big data.” He goes on to explain that machine learning processes when combined with big data, form a sophisticated deep learning environment. What he doesn’t say is that speed isn’t the only benefit machine learning provides to data scientists.
For researchers at Tel Aviv University, machine learning algorithms were able to discern nuances in data that had previously gone undetected. The researchers were interested in understanding the ways bats communicate, so they recorded a group of Egyptian fruit bats for 75 days. According to a report in The Smithsonian Magazine:
Using a modified machine learning algorithm originally designed for recognizing human voices, they fed 15,000 calls into the software.
The researchers discovered that bats were not making random noises. With the help of machine learning, they were able to classify 60 percent of the calls. What is even more interesting is that the calls were almost all argumentative. Bats argue quite a bit, about who they are huddled next to, about food, even about unwanted advances. The call analyses were so nuanced that scientists found that bats use different tones when speaking to other individual bats. This kind of individual communication has been found in very few other species, most notably in humans and dolphins.
Decoding vocalization is a difficult task, but with machine learning algorithms capable of discerning subtle differences, scientists are pulling more meaning from big data.