Animals around humans record millions of acoustic signals that carry information about identity, behavior, environment, and social relationships, but most of these signals are still far from being fully understood. With the development of artificial intelligence , scientists have begun using machine learning algorithms to analyze huge amounts of acoustic recordings and look for patterns that are difficult for humans to detect manually.
This does not mean that machines are now able to translate animal language into human language. Rather, the current role of artificial intelligence is largely to detect, classify, and link sounds to context and behavior, which is a crucial step before attempting to understand what those signals might mean.
Animal communication research faces a huge problem in terms of data volume. A few hours of recordings may contain thousands of different sounds, making manual analysis a lengthy and expensive process. Here, artificial intelligence provides a way to identify, classify, and extract the characteristics of sounds, such as frequency, duration, and sound.The AI system was able to speed up the process of identifying whistles by about 6 times compared to manual analysis. The results also showed that some acoustic characteristics varied depending on the behavior of the dolphins, the presence of fishing nets, and some external sound effects.
These results point to an important benefit of artificial intelligence, which is discovering associations between sound and context, but the existence of a statistical association does not automatically prove that the sound carries a specific meaning similar to words and sentences in human language.
Dolphins are among the animals whose vocal communication patterns have been most studied. It has been known for decades that bottlenose dolphins use so-called signature whistles, a feature that is linked to the identity of certain individuals and can perform a function similar to identifying an individual. Dolphins can also imitate the whistles of other individuals.
Tags:
technology
