AI RESEARCH
Streaming Translation and Transcription Through Speech-to-Text Causal Alignment
arXiv CS.CL
•
ArXi:2603.11578v1 Announce Type: new Simultaneous machine translation (SiMT) has traditionally relied on offline machine translation models coupled with human-engineered heuristics or learned policies. We propose Hikari, a policy-free, fully end-to-end model that performs simultaneous speech-to-text translation and streaming transcription by encoding READ/WRITE decisions into a probabilistic WAIT token mechanism. We also