AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Bioacoustic Geolocation: Species Sounds as Geographic Signals

Mustafa Chasmai, Wuao Liu, Subhransu Maji, Grant Van Horn

ICML 2026 · Vision-language and multimodal › Cross-modal learning · approach: Audio-visual source localization · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
pretrain audio encoder on iNatSounds instead of ImageNet (ablation)AG-CLIP · iNatSounds testregion accuracy: 14.4 -> 17.2Improved
pretrain audio encoder on iNatSounds instead of ImageNet (ablation)AG-CLIP · iNatSounds testcontinent accuracy: 64.9 -> 71.2Improved
add auxiliary checklist loss to GeoCLIP location encoder (ablation)AG-CLIP · iNatSounds testcontinent accuracy: 68.5 -> 71.2Improved
use GeoCLIP location encoder instead of SatCLIP (ablation)AG-CLIP · iNatSounds testregion accuracy: 14.7 -> 17.0Improved
use average pooling over clips instead of max pooling (ablation)AG-CLIP · iNatSounds testregion accuracy: 16.5 -> 17.2Improved
use average pooling instead of K-means cluster pooling (ablation)AG-CLIP · iNatSounds testregion accuracy: 15.5 -> 17.2Improved
use trained transformer pooling instead of averaging (ablation)AG-CLIP · iNatSounds testcontinent accuracy: 71.2 -> 71.5No clear effect
use hierarchical classification instead of single-resolution Res-2 (ablation)Classification · iNatSounds testcountry accuracy: 36.5 -> 40.0Improved
weight SINR maps by all predicted species instead of top-1 species (ablation)Species ranges · iNatSounds testregion accuracy: 08.2 -> 09.3Mixed
use Haversine instead of Euclidean regression loss (ablation)Regression · iNatSounds testregion accuracy: 2.9 -> 4.2Improved
aggregate recordings by Res-5 hexagon over a yearAG-CLIP · iNatSounds testregion accuracy: 17.2 -> 30.4Improved
aggregate recordings by Res-6 hexagon over a week instead of a year (ablation)AG-CLIP · iNatSounds testregion accuracy: 25.9 -> 19.9Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy