Peer-reviewed research

Maximum Entropy Language Modeling for Russian ASR

Evgeniy Shin et al., Proceedings of IWSLT 2013, 2013

Read on October 17, 2026

What it is: A Karlsruhe Institute of Technology team described why Russian is hard for large-vocabulary recognition, because of its inflection, and how a sub-word language model shrank the vocabulary and cut out-of-vocabulary words.
How much weight it carries: Peer-reviewed workshop paper from a research lab. The system is a 2013 design, so the numbers show the problem and the remedy, not the accuracy of modern dictation apps.

Facts taken from this source

Shin and colleagues describe Russian as a challenging language for automatic speech recognition because of its rich morphology.
“Russian is a challenging language for automatic speech recognition systems due to its rich morphology.” (Abstract)
In Shin and colleagues' 2013 Russian system, a sub-word language model cut the out-of-vocabulary rate from 4.8 percent to 2.1 percent.
“the use of the sub-word based language model reduces the vocabulary size from 120k to 40k and the OOV rate from 4.8% to 2.1%” (Abstract)
Shin and colleagues write that Russian nouns can be declined in six cases, with singular and plural numbers and three grammatical genders.
“Russian nouns can be declined according to six cases, two numbers (singular and plural) and three grammatical genders (male, female and neutral).” (Section 1, Introduction)