All projects

Korean STT punctuation restoration research

Improving STT output by restoring punctuation from multimodal signals

ContractTebahSoftSep 2022 – Dec 20224 mosUniversity–industry collaboration

Context

A university–industry project on STT output quality, the front end of the counselling analysis pipeline.

ASR output arrives without punctuation, which hurts readability and drags down every downstream step — summarisation, emotion analysis, and the rest. But audio is not always available in the pipeline, so the model had to use it when present and still work when absent.

How I solved it

  1. Surveyed prior work, confirmed where text-only restoration breaks down, and decided to port UniPunc (ICASSP 2022) to Korean
  2. Encoded text with KoBERT and audio with wav2vec2-xlsr-korean, then aligned the audio frames to text length through a Conv1d subsampler
  3. Fused the two with Transformer cross-attention
    • text as the query, audio as the memory
  4. Substituted a learned virtual embedding wherever audio was missing, so the model trains multimodally but still runs when no audio arrives
  5. Put CLOVA ASR in front of it for an end-to-end Streamlit demo that takes a single audio file

Result

  • Trained and evaluated on AI Hub welfare call-centre counselling data, classifying four marks
    • comma, period, question mark, exclamation mark
  • F1 up 15 points over a text-only KoBERT baseline
  • The same model survives audio going missing in the pipeline
    • it is text-only, so it runs without audio
  • The research was handed over to the company

Tech

PythonPyTorchKoBERTwav2vec2TransformersStreamlit