Research map / Label-efficient and robust learning / Recognition and long-tail
Multi-Instance Multi-Label Class Discovery: A Computational Approach for Assessing Bird Biodiversity
Forrest Briggs, Xiaoli Fern, Raviv Raich, Matthew Betts
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use multi-instance representation methods (cluster centers, MIFF, CCMIFF) instead of dawn time-based sampling | species discovery · H. J. Andrews bird recordings, 100 labeled | species and classes discovered | Improved |
| set instances per bag p=2 and rain threshold T=0.1 in MIFF | MIFF · H. J. Andrews bird recordings, 100 labeled | species and classes discovered | Improved |
| use CCMIFF instead of MIFF when fewer than 50 recordings are labeled | MIFF · H. J. Andrews bird recordings, up to 50 labeled | species discovered | Mixed |
| add rain filter (T=0.1) to MIFF (ablation) | MIFF · H. J. Andrews bird recordings | species and classes discovered | Improved |
| add rain filter (T=0.1) to CCMIFF (ablation) | CCMIFF · H. J. Andrews bird recordings | species and classes discovered | Improved |
| add rain filter to cluster centers (ablation) | cluster centers · H. J. Andrews bird recordings, 100 labeled | classes discovered | Hurt |
| use strictest rain filter T=0.01 for cluster centers at low labeling budgets (ablation) | cluster centers · H. J. Andrews bird recordings, few labeled | classes discovered | Improved |
| lower the rain threshold T (ablation) | all methods · H. J. Andrews bird recordings | number of rain recordings selected | Improved |
| use fewer clusters (k=100) in CCMIFF (ablation) | CCMIFF · H. J. Andrews bird recordings | species discovered | No clear effect |
Closest papers by meaning
- The iNaturalist Sounds Dataset (NeurIPS 2024 · similarity 0.88)
- BirdSet: A Large-Scale Dataset for Audio Classification in Avian Bioacoustics (ICLR 2025 · similarity 0.88)
- NatureLM-audio: an Audio-Language Foundation Model for Bioacoustics (ICLR 2025 · similarity 0.88)
- RealBirdID: Benchmarking Bird Species Identification in the Era of MLLMs (CVPR 2026 · similarity 0.86)
- OpenWhistle: A Large-Scale Longitudinal Dataset and Benchmark of Bottlenose Dolphin Vocalizations (NeurIPS 2026 · similarity 0.84)
- BirdCollect: A Comprehensive Benchmark for Analyzing Dense Bird Flock Attributes (AAAI 2024 · similarity 0.83)
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.
