Update README.md

README.md

@@ -193,7 +193,15 @@ Synthetic training data reduces annotator bias and broadens language coverage, b
 ## Citation
 
 ```bib
-///// comming soon /////
+@misc{borisov2026multilingual,
+      title={Multilingual Multi-Label Emotion Classification at Scale with Synthetic Data}, 
+      author={Vadim Borisov},
+      year={2026},
+      eprint={2604.12633},
+      archivePrefix={arXiv},
+      primaryClass={cs.CL},
+      url={https://arxiv.org/abs/2604.12633}, 
+}
 ```
 
 ## Contact