Automatic Bat Call Classification using Transformer Networks
This addresses the challenge of monitoring bats and ecosystems with improved real-time classification, though it is incremental as it applies an existing Transformer method to a specific domain.
The paper tackles the problem of automatically identifying bat species from echolocation calls by proposing a Transformer architecture for multi-label classification, achieving 88.92% accuracy for single species and 74.40% macro F1-score for multi-species classification, with at least 25.82% better accuracy on an independent dataset compared to other tools.
Automatically identifying bat species from their echolocation calls is a difficult but important task for monitoring bats and the ecosystem they live in. Major challenges in automatic bat call identification are high call variability, similarities between species, interfering calls and lack of annotated data. Many currently available models suffer from relatively poor performance on real-life data due to being trained on single call datasets and, moreover, are often too slow for real-time classification. Here, we propose a Transformer architecture for multi-label classification with potential applications in real-time classification scenarios. We train our model on synthetically generated multi-species recordings by merging multiple bats calls into a single recording with multiple simultaneous calls. Our approach achieves a single species accuracy of 88.92% (F1-score of 84.23%) and a multi species macro F1-score of 74.40% on our test set. In comparison to three other tools on the independent and publicly available dataset ChiroVox, our model achieves at least 25.82% better accuracy for single species classification and at least 6.9% better macro F1-score for multi species classification.