Ecologically-Constrained Task Arithmetic for Multi-Taxa Bioacoustic Classifiers Without Shared Data

2026-05-05

AI Summary

This work addresses the challenges of fragmented and privacy-sensitive bioacoustic data by proposing a collaborative learning framework based on task vector arithmetic. The approach combines independently fine-tuned BEATs encoders into a unified multi-species classifier, requiring only the sharing of task vectors rather than raw audio data. The study reveals that bioacoustic task vectors are approximately orthogonal, with geometric structures aligning with the acoustic niche hypothesis, thereby supporting this privacy-preserving paradigm. Through task vector averaging, linear mode connectivity validation, zero-shot cross-regional transfer, and domain negation boundary analysis, the authors construct a classifier encompassing 661 species, significantly improving recognition accuracy for rare taxa and enabling efficient, scalable biodiversity monitoring.

Abstract

Training data for bioacoustics is scattered across taxa, regions, and institutions. Centralizing it all is often infeasible. We show that independently fine-tuned BEATs encoders can be composed into a unified 661-species classifier via task vector arithmetic without sharing data. We find that bioacoustic task vectors are near-orthogonal (cosine 0.01-0.09). Their separation aligns closely with spectral distribution distance, a gradient consistent with the acoustic niche hypothesis. This geometry makes simple averaging optimal while sign-conflict methods reduce accuracy by one to six percentage points. Composition also creates an asymmetric gap: species-rich groups lose accuracy relative to joint training while underrepresented taxa gain, a redistribution useful for equitable biodiversity monitoring. We verify linear mode connectivity across all taxonomic pairs, demonstrate zero-shot transfer to new regions, and identify domain negation as a boundary condition where composition fails. These results enable a collaborative paradigm for bioacoustics where institutions share only task vectors to assemble multi-taxa classifiers, preserving data privacy.

Problem

• bioacoustics
• multi-taxa classification
• data privacy
• distributed training data
• species classification

Innovation

• task vector arithmetic
• bioacoustic classification
• data privacy
• acoustic niche hypothesis
• linear mode connectivity