# Efficient ASR Training with Conversations that Never Happened

arXiv: https://arxiv.org/abs/2606.03957 · HTML: https://mlsanity.com/paper/2606.03957/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 28 of its 47 references that are in the map.
- First posted on arXiv 2 Jun 2026.

## Builds on (most cited first, 28 of 28)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://mlsanity.com/paper/2006.11477/) (2020)
- [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://mlsanity.com/paper/2106.07447/) (2021)
- [Conformer: Convolution-augmented Transformer for Speech Recognition](https://mlsanity.com/paper/2005.08100/) (2020)
- [WaveNet: A Generative Model for Raw Audio](https://mlsanity.com/paper/1609.03499/) (2016)
- [SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition](https://mlsanity.com/paper/1904.08779/) (2019)
- [Natural TTS Synthesis by Conditioning WaveNet on Mel Spectrogram Predictions](https://mlsanity.com/paper/1712.05884/) (2017)
- [Listen, Attend and Spell](https://mlsanity.com/paper/1508.01211/) (2015)
- [Tacotron: Towards End-to-End Speech Synthesis](https://mlsanity.com/paper/1703.10135/) (2017)
- [NeMo: a toolkit for building AI applications using Neural Modules](https://mlsanity.com/paper/1909.09577/) (2019)
- [XTTS: a Massively Multilingual Zero-Shot Text-to-Speech Model](https://mlsanity.com/paper/2406.04904/) (2024)
- [CHiME-6 Challenge:Tackling Multispeaker Speech Recognition for Unsegmented Recordings](https://mlsanity.com/paper/2004.09249/) (2020)
- [Fast Conformer with Linearly Scalable Attention for Efficient Speech Recognition](https://mlsanity.com/paper/2305.05084/) (2023)
- [End-to-End Neural Speaker Diarization with Permutation-Free Objectives](https://mlsanity.com/paper/1909.05952/) (2019)
- [Conversational End-to-End TTS for Voice Agent](https://mlsanity.com/paper/2005.10438/) (2020)
- [FireRedTTS-2: Towards Long Conversational Speech Generation for Podcast and Chatbot](https://mlsanity.com/paper/2509.02020/) (2025)
- [From Simulated Mixtures to Simulated Conversations as Training Data for End-to-End Neural Diarization](https://mlsanity.com/paper/2204.00890/) (2022)
- [Generating Data with Text-to-Speech and Large-Language Models for Conversational Speech Recognition](https://mlsanity.com/paper/2408.09215/) (2024)
- [Improving the Naturalness of Simulated Conversations for End-to-End Neural Diarization](https://mlsanity.com/paper/2204.11232/) (2022)
- [Pheme: Efficient and Conversational Speech Generation](https://mlsanity.com/paper/2401.02839/) (2024)
- [From Independence to Interaction: Speaker-Aware Simulation of Multi-Speaker Conversational Timing](https://mlsanity.com/paper/2509.15808/) (2025)
- [Toward Conversational Hungarian Speech Recognition: Introducing the BEA-Large and BEA-Dialogue Datasets](https://mlsanity.com/paper/2511.13529/) (2025)
- [Speaker-Aware Simulation Improves Conversational Speech Recognition](https://mlsanity.com/paper/2602.04776/) (2026)
- [Scaling Conversational Hungarian ASR: The BEA-Dialogue+ Corpus](https://mlsanity.com/paper/2605.31469/) (2026)

## Abstract (arXiv)

Conversational ASR for lower-resource languages and niche domains is limited by the scarcity of domain-matched multi-speaker training data. We propose an augmentation pipeline that generates scenario-level dialogues with participant metadata, maps speaker attributes to TTS voice profiles, and assembles synthesized utterances into speaker-aware simulated conversations. We evaluated five LLM families under single-generator, fixed-budget mixture, and scale-up settings using the same FastConformer-Large training recipe for each one. We ran comprehensive evaluations on the Hungarian BEA-Dialogue benchmark corpus, with the method itself being applicable to any language given the resources for each component. The results show that synthetic conversations consistently improve speech recognition performance, but generator choice and data composition strongly affect the gains. Our largest training configuration, using only 67 hours of real conversations and 636 hours of simulated data, achieves better performance on the evaluation benchmark than a zero-shot model trained on 2700 hours of Hungarian speech. These findings indicate that LLM-generated conversational data synthesized with TTS is a practical complement to real conversational corpora for speech model training.