What we have been reading: recommended papers in Bioacoustic AI

I asked our 10 doctoral candidates for their favourite research papers they’ve been reading in 2024. Their responses were really interesting – altogether, a great list of research papers. Inspiration for bioacoustic AI! Here are their papers of the year for 2024, each with a brief comment from a doctoral candidate. Maybe you can find an inspiring paper from this list!

• Acoustic monitoring for tropical insect conservation (Riede & Balakrishnan 2024)
  “Very interesting and well written. It makes a good case for why insect monitoring is important, but also gives a very good overview of the sonic variation and sound production methods of insects and outlines further work that has to be done to implement bioacoustic monitoring.”
• Acoustic localization of terrestrial wildlife: Current practices and future opportunities (Rhinehart et al 2020)
  “In acousting wildlife monitoring, there are tons of small (but critical) details that we have to consider during the processes of planning, preparing and deploying the recorders … This review helped me a lot on that, after reading it, I felt like all the other bioacoustic/wildlife monitoring articles that I read make more sense.”
• An environmental assessment method for wireless sensor networks (Bonvoisin et al 2012)
  “I read this paper around May but I still think about it because its environmental impact assessment method accounts for changes during deployment, making it interesting (in my opinion) for evaluating autonomy along with sustainability.”
• NatureLM-audio: an Audio-Language Foundation Model for Bioacoustics (Robinson et al 2024)
  “Very inspiring in the direction of creating LLMs for Bioacoustics. This paper started with the first paper by Robinson on LLMs in Bioacoustics that uses a clever idea of creating a dataset pairing metadata/text with audio to create a multimodal-language model to use in different bioacoustics tasks.”
• Global cultural evolutionary model of humpback whale song (Zandberg et al 2021)
  “This paper is about the global pattern of song culture transmissions of humpback whales. Due to geographic differences of population movements, it shows how the dynamics of song culture are different between the northern and southern hemispheres through simulations. It was fascinating to me how song culture dynamics can be identified in such large scale with modelling methodologies.”
• A densely sampled and richly annotated acoustic data set from a wild bird population (Recalde et al 2024)
  “This is a very recent paper based on a great tit song dataset and a song analysis package ‘pykanto’. The research question is very much similar to my current work, and it gave me a lot of inspiration.”
• Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hypersphere (Wang & Isola 2020)
  “It got me thinking about deep embeddings in a different way.”
• Birds, bats and beyond: evaluating generalization in bioacoustics models (van Merrienboer et al 2024)
  “It gives a thorough overview of the process of evaluating bioacoustics models and the associated challenges.”