birdclef

Species Identification from Bioacoustics

Solution for BirdCLEF+ 2025

Code and experiments for my submission to the BirdCLEF+ 2025 Kaggle competition.

Demonstrating the "Reading the Robot Mind" system using a bird audio recognition AI system.

基于 BirdCLEF 2026 的声音鸟类识别课程项目，包含语音特征工程、传统机器学习、CNN/CRNN 对比、GroupKFold 泛化分析与 Gradio 可视化演示。

Solo late-entry BirdCLEF+ 2026 solution + CLEF working note on ensemble diversity

Animal sound classifier for BirdCLEF+ 2025 — EfficientNet B0 + FastAPI, with a Gemini agent that checks whether detections are plausible given your context.

Design of an ablation study for a Machine Learning pipeline. The effect of preprocessing, model or postprocessing modules can be automatically tested.

Bird sound recognition & language system — identify species and call types from audio (CLI + Gradio GUI)

Classical ML pipeline for BirdCLEF+ 2026: identifying 206 wildlife species from Pantanal audio recordings with hand-engineered features, no neural networks. 0.922 macro ROC-AUC.

BirdCLEF+ 2026 (Kaggle): multi-label species detection in audio soundscapes with handcrafted features and LightGBM — leakage-free grouped validation, co-occurrence correction, soundscape sub-window features (public macro-AUC 0.783)

Ablation study for BirdCLEF+ 2026 — companion code for Kaggle notebook series and CLEF 2026 working note

I made this project with 2 fellow students as a solution for the 2024 BirdCLEF competition.

Bioacoustic audio classification and soundscape analysis using deep learning and ensemble methods.

BirdCLEF+ 2026 bioacoustic detection with Perch embeddings, ProtoSSM, OOF stacking and TTA

BirdCLEF 2024

BirdCLEF+ 2025 multi-label bird sound event detection — versioned PyTorch SED pipeline (v3 val macro ROC-AUC 0.9694).

Acoustic species ID on BirdCLEF 2026 — BirdNET embeddings + ML vs. from-scratch CNN, testing AudioLDM2 synthetic-data enrichment (0.894 macro ROC-AUC).