This research paper presents an innovative approach to anomaly detection in maritime data, focusing on the identification of anomalous vessel groups through Automatic Identification System (AIS) data. Traditional AIS anomaly detection systems are challenged by a narrow focus on individual vessel anomalies which leaves out critical collective action information. Our work seeks to overcome this limitation by introducing a graph modeling methodology that emphasizes the collective behavior of vessels, a critical aspect often overlooked in current frameworks. We work towards a system that detects unusual patterns of coordination among groups of vessels, which is particularly relevant for defense contexts where threats such as smuggling, piracy, and adversarial operations are typically carried out by networks of collaborating entities. The research contributes to the field by offering a more comprehensive understanding of maritime traffic through the lens of collective dynamics, enhancing the detection capabilities of AIS monitoring systems. The implications of our methodology are far-reaching, providing a foundational strategy for future research in anomaly detection that could be applied to various domains where group coordination plays a pivotal role in defining anomalous behavior. This paper details our graph-based approach and suggests a trajectory for subsequent investigations into the broader applications of detecting coordinated anomalies.

Wildlife biologists often rely on classification models to automate the labelling process of audio data that they collect in the wild. Bird populations are often a subject of interest in audio data, and tens to hundreds of species can be tracked in population-dense areas such as the Amazon rainforest. However, these population-dense environments also pose the issue of heavy noise, distortion, and overlapping bird calls to be classified. By transforming audio signals into their spectrogram form in the frequency domain, we can identify keypoints in the spectrogram signal by passing a sliding window over the signal and finding the local maxima. Describing an audio signal by its keypoint-spectrogram can result in higher classification accuracy for overlapping bird calls when passed as input to a CNN than if the entirety of the audio spectrogram were to be used.