Toward understanding the communication in sperm whales

Area of Science:

• Bioacoustics
• Machine Learning
• Animal Communication

Background:

• Machine learning (ML) advancements have primarily focused on human applications due to large datasets.
• Significant opportunities exist to apply ML for understanding non-human communication systems.

Purpose of the Study:

• To present a scientific roadmap for advancing the understanding of whale communication using ML.
• To establish a template for deciphering other animal communication forms.

Main Methods:

• Utilizing sperm whales as a model due to their complex communication and neuroanatomy.
• Outlining requirements for massive dataset collection and processing.
• Developing methods for detecting communication units and language-like structures.
• Validating ML models through interactive playback experiments.

Main Results:

• The proposed roadmap details the essential steps for analyzing complex animal vocalizations.
• It highlights the potential for identifying discrete communication units and higher-level structures.
• The framework is designed for scalability and application to diverse species.

Conclusions:

• The developed technological capabilities can significantly advance non-human communication research.
• This approach offers a powerful tool for investigating animal behavior and cognition.
• The proposed framework serves as a foundational template for future interspecies communication studies.