[KongchangAI]
Conceptcontextual biasing

Contextual Biasing

A technique in speech recognition that injects hot words such as contact names and proper nouns into the decoding graph to improve recognition accuracy for specific vocabulary.

Timeline (last 90 days)

Nov 16

Stronger contextual bias does not imply a higher neuron-feature verification rate; behavioral-level context dependence and the semantic interpretability of hidden neurons may capture different aspects of CNN behavior.

Unverified50%
Nov 10

Contextual biasing is a technique that introduces an external vocabulary or phrase list at the decoding stage to increase the recognition probability of specific word sequences by adjusting language model scores or using methods such as shallow fusion.

Unverified50%
Nov 10

A hotword list is a simplified form of contextual biasing that typically contains only a list of keywords without distinguishing weights.

Unverified50%
Nov 3

In speech recognition research, contextual biasing and prompt-based shallow fusion are existing concepts, the former injecting hotwords into the decoding graph and the latter encoding preceding text into vectors that participate in acoustic model scoring.

Unverified50%

All Facts (4)

Source Articles