Enhance README with a new section detailing the research paper on the hybrid tokenization approach, including citation information and authors. Update requirements to specify the version of the Turkish Tokenizer package.

README.md

@@ -55,6 +55,30 @@ This tokenizer uses:
 - Gradio for the web interface
 - Advanced tokenization algorithms
 
+## Research Paper
+
+This implementation is based on the research paper:
+
+**"Tokens with Meaning: A Hybrid Tokenization Approach for NLP"**
+
+📄 [arXiv:2508.14292](https://arxiv.org/abs/2508.14292)
+
+**Authors:** M. Ali Bayram, Ali Arda Fincan, Ahmet Semih Gümüş, Sercan Karakaş, Banu Diri, Savaş Yıldırım, Demircan Çelik
+
+**Abstract:** A hybrid tokenization framework combining rule-based morphological analysis with statistical subword segmentation improves tokenization for morphologically rich languages like Turkish. The method uses phonological normalization, root-affix dictionaries, and a novel algorithm that balances morpheme preservation with vocabulary efficiency.
+
+Please cite this paper if you use this tokenizer in your research:
+
+```bibtex
+@article{bayram2024tokens,
+  title={Tokens with Meaning: A Hybrid Tokenization Approach for NLP},
+  author={Bayram, M. Ali and Fincan, Ali Arda and Gümüş, Ahmet Semih and Karakaş, Sercan and Diri, Banu and Yıldırım, Savaş and Çelik, Demircan},
+  journal={arXiv preprint arXiv:2508.14292},
+  year={2025},
+  url={https://arxiv.org/abs/2508.14292}
+}
+```
+
 ## Files
 
 - `app.py`: Main Gradio application

requirements.txt

@@ -1,2 +1,2 @@
 gradio
-turkish-tokenizer
+turkish-tokenizer==0.2.24