Do Elephants Have Names? Vocal Labeling of Individual Conspecifics in African Elephants + ...
summary
Wild African elephant rumbles analyzed from Samburu and Amboseli (625 contact and greeting calls) show acoustic signatures that identify intended receivers: a random forest trained on PCA-derived spectrogram features classifies receiver above chance (~20.3% accuracy versus randomized nulls) while proximity analyses control for caller and relationship. Different callers partially converge on similar labels, calls are not simple imitations of receivers, and a playback experiment (17 subjects) elicits faster approach, quicker calling, and more calls for self-addressed rumbles.
AI-written summary · about
chapters (13)
- 0:00Introduction to the study
- 0:40Defining vocal labeling
- 3:24Elephant communication types
- 6:16Research objectives and methods
- 8:55Acoustic feature extraction
- 12:54Random forest classification
- 16:47Ruling out alternative theories
- 21:30Testing across callers
- 27:13Imitation vs. arbitrary names
- 28:43Playback experiment results
- 32:15Conclusions and implications
- 34:25Future research directions
- 36:41Q&A with Joyce Poole
related talks
56mWhistle variability and social acoustic interactions in bottlenose dolphins
43mDeciphering animal emotions using machine learning and vocalisations
25mThe Sound of Life
45mEvidence of social learning across symbolic cultural barriers in sperm whales
53mAI as a tool for discovery in animal communication
15mThe Internet of Wild Things: Acoustics, AI and Conservation
01h 07mDecoding Dolphin Speech
54mPotential of Echolocation for Communication