cs.SDcs.LGeess.ASq-bio.QMEmpirical

Adaptive Loss Balancing for Multi-Task Bioacoustic Classification of Bird Species and Call Types

Jul 3, 2026

AI Summaryllama-3.1-8b-instruct

This paper extends BirdCallNet for joint species and call-type classification on the long-tailed WiWa dataset and investigates the interaction of task-loss balancing with pretrained representations and adaptation depth.

Investigates the interaction of task-loss balancing with pretrained representations and adaptation depth for bird vocalization analysis

Keywords

bird vocalizations passive acoustic monitoring multi-task learning pretrained representations adaptation depth

Before reading this…

Understanding of multi-task learning and pretrained representations Familiarity with bird vocalization analysis and passive acoustic monitoring

Applications

→Bird conservation
→Biodiversity monitoring

Skill Ladder

To understand this paper, make sure you know these concepts first:

Understanding of multi-task learning and pretrained representationsfind papers →
Familiarity with bird vocalization analysis and passive acoustic monitoringfind papers →

Abstract

More Like This

Reliable analysis of bird vocalisations in passive acoustic monitoring requires models handling multiple, imbalanced annotation targets. We extend BirdCallNet for joint species and call-type classification on the long-tailed WiWa dataset and investigate how task-loss balancing interacts with pretrained representations and adaptation depth. We evaluate four bird-domain encoders, ConvNeXtBS, EAT, BirdMAE, and ProtoCLR, with separate species and call-type heads under linear probing, attentive probing, and full fine-tuning. A manually tuned fixed objective is compared with homoscedastic uncertainty weighting and Dynamic Weight Averaging across all three adaptation regimes, while GradNorm is evaluated only under full fine-tuning. Results indicate that the factorised multi-task formulation yields the most consistent improvements over the combined single-task baseline for call-type recognition, while its effect on species recognition depends on the adaptation regime. Full fine-tuning is not consistently optimal: ConvNeXtBS achieves the highest mean species performance under linear probing, whereas BirdMAE provides the strongest call-type performance under attentive probing. Adaptive weighting benefits species recognition more consistently than call-type recognition. Uncertainty weighting is particularly effective for species recognition under attentive probing, whereas Dynamic Weight Averaging is generally stronger for the same task under full fine-tuning. GradNorm achieves competitive call-type performance for selected backbones but consistently underperforms other weighting strategies for species recognition and incurs higher computational and memory costs. Overall, the preferred loss-balancing strategy depends on the backbone, adaptation regime, and target task, while frozen-backbone adaptation can provide a more favourable performance-efficiency trade-off than end-to-end fine-tuning.