docs: add INTERSPEECH 2026 paper banner, badges, and citation

README.md

@@ -8,6 +8,16 @@ pipeline_tag: audio-classification
 ---
 # Music Detection with WavLM
 
+[![arXiv](https://img.shields.io/badge/arXiv-2507.13563-b31b1b.svg)](https://arxiv.org/abs/2507.13563)
+[![Conference](https://img.shields.io/badge/INTERSPEECH-2026-1f6feb.svg)](https://arxiv.org/abs/2507.13563)
+[![Code](https://img.shields.io/badge/code-lab260ru%2Fbalalaika-181717.svg?logo=github)](https://github.com/lab260ru/balalaika)
+
+> [!IMPORTANT]
+> **Official model for our INTERSPEECH 2026 paper**
+> *"A Data-Centric Framework for Addressing Phonetic and Prosodic Challenges in Russian Speech Generative Models"* ([arXiv:2507.13563](https://arxiv.org/abs/2507.13563)).
+> Part of the **Balalaika** Russian speech data-processing pipeline — code: [https://github.com/lab260ru/balalaika](https://github.com/lab260ru/balalaika).
+> If you use this resource, please [cite it](#citation).
+
 Detects if audio contains music.  
 **EER: 2.5–3%** | Based on `microsoft/wavlm-base-plus`
 *the best threshold value* `0.2442`
@@ -34,3 +44,18 @@ probs = model.predict_proba(['audio1.mp3', 'audio2.wav'])  # → tensor([0.88, 0
 
 - Email: kborodin.research@gmail.com
 - Telegram: [@korallll_ai](https://t.me/korallll_ai)
+
+## Citation
+
+If you use this resource, please cite our INTERSPEECH 2026 paper:
+
+```bibtex
+@inproceedings{borodin2026balalaika,
+  title     = {A Data-Centric Framework for Addressing Phonetic and Prosodic Challenges in Russian Speech Generative Models},
+  author    = {Borodin, Kirill and Vasiliev, Nikita and Kudryavtsev, Vasiliy and Maslov, Maxim and Gorodnichev, Mikhail and Rogov, Oleg and Mkrtchian, Grach},
+  booktitle = {Proc. INTERSPEECH 2026},
+  year      = {2026},
+  note      = {arXiv:2507.13563},
+  url       = {https://arxiv.org/abs/2507.13563}
+}
+```