Understanding the communication of whales is important not just for expanding our knowledge, but also for protecting this endangered species. LTU/MCS is now taking part in a large international project that aims at profiling the way whales communicate. The project includes several years of audio data acquired by using special audio sensors attached to the whales, and while watching the whales in their natural environment from ships. LTU’s part in the project is to develop pattern recognition and machine learning algorithms, and apply these algorithms to mine hours of data recordings and tens of thousands of whale communication recording samples. Our machine learning methods are trained with the help of over 800,000 volunteers who annotate the data manually. Our first findings have been accepted for publication in the prestigious Journal of the Acoustical Society of America
In the project LTU is part of a collaboration that includes researchers from Woods Hole Oceanographic Institute, St. Andrews University, Oxford University (UK) and TNO (Netherlands). The project also involves LTU student Carol Yerby, who is one of the authors of the paper.