Streaming Automatic Speech Recognition
Definition
Streaming speech recognition continuously converts ongoing speech into initially provisional text.
Streaming speech recognition processes audio during recording and continuously delivers provisional as well as later stabilized transcripts.
Simply explained
Subtitles appear while you speak. When more of the sentence is known, the system can still correct early words.
Why it matters in practice
The term helps to select, develop, evaluate AI systems in a well-founded manner and realistically assess their technical risks.
Example from business
A live captioning service shows interim results and corrects them as more context becomes available.
Limits and pitfalls
- Early partial hypotheses can change; latency, accuracy, and privacy must be evaluated together.
Origin & history
Online decoders and neural transducers enabled low latency in continuous speech recognition.
Use in mid-sized companies
- Possible application: A live captioning service shows interim results and corrects them as more context becomes available.
Frequently asked questions
What does streaming speech recognition mean?
Streaming speech recognition processes audio during recording and continuously delivers provisional as well as later stabilized transcripts. What should be considered? Early partial hypotheses can change; latency, accuracy, and privacy must be evaluated together.
Sources
- Sequence Transduction with Recurrent Neural Networks · Graves, Sequence Transduction with Recurrent Neural Networks
Source checked on 10.10.2026. Editorial use cases are hypothetical; no evidence of real customer results.
Editorially reviewed on: