1 paper
Zhaohui Liang, Yonglin Chen, Naser Al Madi +1
Transcripts displayed on dictation interfaces can be hard to read due to recognition errors and disfluencies. LLM-based text auto-correction could help, but changing the text durin…