Vocalization, Animal

Florida manatees are semisocial marine mammals that vocalize when interacting with conspecifics and to maintain contact with offspring. While many aspects of their biology have been studied, there is a dearth of information on the diversity and complexity of vocal behavior during social, nonsocial, and stressful situations. Investigations of vocal communication repertoires which define, categorize, and correlate varied call types with behavior are needed in order to understand the behavioral and social function of associated calls. Arguably the most important social bond in manatees is the period of cow/calf dependency and empirical evidence indicates cows recognize the vocalizations of offspring. Exploration of individually distinctive vocal features can provide insight on which parameters might be salient to facilitate recognition between cows/calves. This study is focused on vocal communication in Florida manatees, how calls are structured, utilized and function while animals are distressed and during social interactions in their shallow water habitats. Hydrophones recorded vocalizations from individual calves and manatees in different behavioral contexts and varying size aggregations. Analysis of the vocal repertoire indicated manatee vocalizations can be parsed into five broadly defined call types which include the hill-shaped high squeak, tonal squeak, noisy squeal, two toned chirp, and the combinatorial squeak-squeal. Furthermore, the high squeak is likely a discrete call whereas the others are graded and do not have strict boundaries between call types (Chapter 2). Broadly defined call types were used to explore call usage with variations in behavior, group size, and group composition (Chapter 3). Manatees vocalized using few call types and altered structural parameters depending on behavioral state. Calls were longer and more frequency modulated when stressed. Vocalizations produced while cavorting were higher in entropy and more frequency modulated than when manatees were resting or feeding. Vocalizations obtained from individual calves suggest that the high squeak is a stereotypical call that is produced by smaller calves. All calves had individually distinctive acoustic features that could potentially be used in recognition (Chapter 4). Lower fundamental frequencies and higher emphasized frequencies from smaller calves suggest that the fundamental frequency may not be a reliable indicator of body size in calves. This research increases our knowledge of the vocal behavior and call characteristics of the Florida manatee.

Deep learning models have been successfully applied to a variety of machine learning tasks, including image identification, image segmentation, object detection, speaker recognition, natural language processing, bioinformatics and drug discovery, among other things. This dissertation introduces Multi-Model Deep Learning (MMDL), a new ensemble deep learning approach for signal classification and event forecasting. The ultimate goal of the MMDL method is to improve classification and forecasting performances of individual classifiers by fusing results of participating deep learning models. The performance of such an ensemble model, however, depends heavily on the following two design features. Firstly, the diversity of the participating (or base) deep learning models is crucial. If all base deep learning models produce similar classification results, then combining these results will not provide much improvement. Thus, diversity is considered to be a key design feature of any successful MMDL system. Secondly, the selection of a fusion function, namely, a suitable function to integrate the results of all the base models, is important. In short, building an effective MMDL system is a complex and challenging process which requires deep knowledge of the problem context and a well-defined prediction process. The proposed MMDL method utilizes a bank of Convolutional Neural Networks (CNNs) and Stacked AutoEncoders (SAEs). To reduce the design complexity, a randomized generation process is applied to assign values to hyperparameters of base models. To speed up the training process, new feature extraction procedures which captures time-spatial characteristics of input signals are also explored. The effectiveness of the MMDL method is validated in this dissertation study with three real-world case studies. In the first case study, the MMDL model is applied to classify call types of groupers, an important fishery resource in the Caribbean that produces sounds associated with reproductive behaviors during yearly spawning aggregations. In the second case study, the MMDL model is applied to detect upcalls of North Atlantic Right Whales (NARWs), a type of endangered whales. NARWs use upcalls to communicate among themselves. In the third case study, the MMDL model is modified to predict seizure episodes. In all these cases, the proposed MMDL model outperforms existing state-of-the-art methods.

Chimpanzees have long been documented as using population-specific
vocalizations, implying learning rather than just genetics in chimpanzee calls. In order
for population-specific vocalizations to arise, diachronic change, or evolution, of the
various features of the vocalizations must occur. When a population is split, as they were
in the current study, there are changes of social structure, environment, and emotional
stress (all factors which can lead to rapid phonological change in humans). These factors
can act as a catalyst for punctuated diachronic change. A vocal survey was performed on
two groups of chimpanzees who had been separated from each other two years prior to
the research. The results of the survey revealed significant differences between the two
groups' vocalizations. These results make a case for diachronic change in chimpanzee
vocalizations, the seed of population-specific calls.