# Optimizing Estonian TV Subtitles with Semi-supervised Learning and LLMs

arXiv: https://arxiv.org/abs/2501.05234 · HTML: https://mlsanity.com/paper/2501.05234/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 11 of its 22 references that are in the map.
- First posted on arXiv 9 Jan 2025.

## Builds on (most cited first, 11 of 11)

- [Optuna: A Next-generation Hyperparameter Optimization Framework](https://mlsanity.com/paper/1907.10902/) (2019)
- [SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition](https://mlsanity.com/paper/1904.08779/) (2019)
- [Averaging Weights Leads to Wider Optima and Better Generalization](https://mlsanity.com/paper/1803.05407/) (2018)
- [BLEURT: Learning Robust Metrics for Text Generation](https://mlsanity.com/paper/2004.04696/) (2020)
- [Learning Compact Metrics for MT](https://mlsanity.com/paper/2110.06341/) (2021)
- [Iterative Pseudo-Labeling for Speech Recognition](https://mlsanity.com/paper/2005.09267/) (2020)
- [Can Generative Large Language Models Perform ASR Error Correction?](https://mlsanity.com/paper/2307.04172/) (2023)
- [Multi-stage Large Language Model Correction for Speech Recognition](https://mlsanity.com/paper/2310.11532/) (2023)
- [Retrieval Augmented Correction of Named Entity Speech Recognition Errors](https://mlsanity.com/paper/2409.06062/) (2024)
- [SubER: A Metric for Automatic Evaluation of Subtitle Quality](https://mlsanity.com/paper/2205.05805/) (2022)
- [Semi-supervised Learning for Code-Switching ASR with Large Language Model Filter](https://mlsanity.com/paper/2407.04219/) (2024)

## Built on by (most cited first, 1 of 1)

- [From Speech to Subtitles: Evaluating ASR Models in Subtitling Italian Television Programs](https://mlsanity.com/paper/2512.19161/) (2025)

## Abstract (arXiv)

This paper presents an approach for generating high-quality, same-language subtitles for Estonian TV content. We fine-tune the Whisper model on human-generated Estonian subtitles and enhance it with iterative pseudo-labeling and large language model (LLM) based post-editing. Our experiments demonstrate notable subtitle quality improvement through pseudo-labeling with an unlabeled dataset. We find that applying LLM-based editing at test time enhances subtitle accuracy, while its use during training does not yield further gains. This approach holds promise for creating subtitle quality close to human standard and could be extended to real-time applications.