# Post-decoder Biasing for End-to-End Speech Recognition of Multi-turn Medical Interview

arXiv: https://arxiv.org/abs/2403.00370 · HTML: https://mlsanity.com/paper/2403.00370/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 40 references that are in the map.
- First posted on arXiv 1 Mar 2024.

## Builds on (most cited first, 22 of 22)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Sequence to Sequence Learning with Neural Networks](https://mlsanity.com/paper/1409.3215/) (2014)
- [SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing](https://mlsanity.com/paper/1808.06226/) (2018)
- [Conformer: Convolution-augmented Transformer for Speech Recognition](https://mlsanity.com/paper/2005.08100/) (2020)
- [SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition](https://mlsanity.com/paper/1904.08779/) (2019)
- [ESPnet: End-to-End Speech Processing Toolkit](https://mlsanity.com/paper/1804.00015/) (2018)
- [Listen, Attend and Spell](https://mlsanity.com/paper/1508.01211/) (2015)
- [GigaSpeech: An Evolving, Multi-domain ASR Corpus with 10,000 Hours of Transcribed Audio](https://mlsanity.com/paper/2106.06909/) (2021)
- [Deep context: end-to-end contextual speech recognition](https://mlsanity.com/paper/1808.02480/) (2018)
- [An analysis of incorporating an external language model into a sequence-to-sequence model](https://mlsanity.com/paper/1712.01996/) (2017)
- [Contextualized Streaming End-to-End Speech Recognition with Trie-Based Deep Biasing and Shallow Fusion](https://mlsanity.com/paper/2104.02194/) (2021)
- [Knowledge-Enriched Transformer for Emotion Detection in Textual Conversations](https://mlsanity.com/paper/1909.10681/) (2019)
- [Deep Shallow Fusion for RNN-T Personalization](https://mlsanity.com/paper/2011.07754/) (2020)
- [Earnings-21: A Practical Benchmark for ASR in the Wild](https://mlsanity.com/paper/2104.11348/) (2021)
- [Improving End-to-End Contextual Speech Recognition with Fine-Grained Contextual Knowledge Selection](https://mlsanity.com/paper/2201.12806/) (2022)
- [Contextual Speech Recognition with Difficult Negative Training Examples](https://mlsanity.com/paper/1810.12170/) (2018)
- [Dialog-context aware end-to-end speech recognition](https://mlsanity.com/paper/1808.02171/) (2018)
- [Speech recognition for medical conversations](https://mlsanity.com/paper/1711.07274/) (2017)
- [Improving Contextual Recognition of Rare Words with an Alternate Spelling Prediction Model](https://mlsanity.com/paper/2209.01250/) (2022)
- [End-to-end contextual asr based on posterior distribution adaptation for hybrid ctc/attention system](https://mlsanity.com/paper/2202.09003/) (2022)
- [Improved Factorized Neural Transducer Model For text-only Domain Adaptation](https://mlsanity.com/paper/2309.09524/) (2023)
- [Evaluating Automatic Speech Recognition in an Incremental Setting](https://mlsanity.com/paper/2302.12049/) (2023)

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

- [Distilling Conversations: Abstract Compression of Conversational Audio Context for LLM-based ASR](https://mlsanity.com/paper/2603.26246/) (2026)

## Abstract (arXiv)

End-to-end (E2E) approach is gradually replacing hybrid models for automatic speech recognition (ASR) tasks. However, the optimization of E2E models lacks an intuitive method for handling decoding shifts, especially in scenarios with a large number of domain-specific rare words that hold specific important meanings. Furthermore, the absence of knowledge-intensive speech datasets in academia has been a significant limiting factor, and the commonly used speech corpora exhibit significant disparities with realistic conversation. To address these challenges, we present Medical Interview (MED-IT), a multi-turn consultation speech dataset that contains a substantial number of knowledge-intensive named entities. We also explore methods to enhance the recognition performance of rare words for E2E models. We propose a novel approach, post-decoder biasing, which constructs a transform probability matrix based on the distribution of training transcriptions. This guides the model to prioritize recognizing words in the biasing list. In our experiments, for subsets of rare words appearing in the training speech between 10 and 20 times, and between 1 and 5 times, the proposed method achieves a relative improvement of 9.3% and 5.1%, respectively.