Course details

8.3479

Machine Learning meets Animal communication

 SP
WS 2020 Ulf Krumnack, Prof. Dr. Simone Pika, Axel Schaffland, Rachael Xi Cheng ONLINE
2h/wk
12 ECTS 		M.Sc modules:
CC-MP-SP - Study Project
CS-MP-SP - Study Project
 Tue: 12-13

Machine Learning Meets Animal Communication There has been a long-standing interest to understand the meaning and function of animal vocalizations, as well as the structures which determine how animals communicate (Darwin, 1871). In human languages, a phoneme is the smallest meaningful unit of sound and syntactic rules governing sentence formation and sematic rules guiding assignment of meanings to sentences play a crucial role. However, what are the smallest components of complex animal vocalizations? According to which underlying principles are they organized within the respective communication systems to make the communication intelligible? What can we infer about their underlying cognitive processes from those underlying rules? The majority of comparative research into language origins has been based on traditional methods such as behavioural observations and play-back experiments (Seyfarth et al. 1980). The large amount of data and variations in the vocalizations as well as subjectivity make hand analysis challenging. Recently, machine learning has advanced many different scientific fields that impact our daily lives, such as automated speech recognition and computer vision. Its application is, however, still extremely underdeveloped in the fields of Animal Behaviour, Bioacoustics, and Comparative Psychology. The goals and the work packages of the study project The interdisciplinary study project of Machine Learning Meets Animal Communication aims to fill this gap by developing parts of a user-friendly framework that supports the extraction of detectable principles within the vocal repertoire of one of our closest living relatives, the chimpanzee (Pan troglodytes): The project is divided in 3 working packages, 2 deliverables and will be assessed via presentations and a final report. (The following working packages and deliverables serve as a guidance to the topic. Participating students can adjust and develop the final focus ...