STREAMING RUSSIAN-UZBEK TEXT TRANSLATION UNDER CODE-SWITCHING CONSTRAINTS
Keywords:
Streaming translation, Russian-Uzbek, code-switching, wait-k, monotonic attention, morphology-aware tokenization, Average Lagging, chat translation.Abstract
In this article, we investigate streaming Russian-Uzbek machine translation under dense, spontaneous code-switching. We propose a latency-controlled Transformer that combines wait-k scheduling with monotonic chunkwise attention, augmented by script and morphology-aware tokenization and a boundary-sensitive read/write policy. On chat-style test sets, the system delivers higher BLEU and chrF at the same Average Lagging and better preserves switch points and Russian stems bearing Uzbek suffixes.
Downloads
Published
Issue
Section
License

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.











