Exclusive: Speech recognition AI learns industry jargon with aiOla's novel approach
What key advancement did aiOla announce in speech recognition?
aiOla announced a major milestone in speech recognition with its solution achieving human proficiency in understanding industry-specific jargon. The patented AdaKWS model reached 95% accuracy in keyword spotting, surpassing OpenAI's industry-leading Whisper model which reached 88% accuracy6.
How does aiOla's technique improve Whisper's model?
aiOla's technique improves Whisper's model by using a two-step "contextual biasing" approach. First, their AdaKWS keyword spotting model identifies domain-specific and personalized jargon. Then, these identified keywords are utilized to prompt the ASR decoder, guiding it to incorporate them into the final transcribed text. This enhances the model's speech recognition capability, adapting it to correctly detect jargon or specific terms3.
What industries can benefit from aiOla's contextual biasing approach?
aiOla's contextual biasing approach can benefit industries such as aviation, transportation, manufacturing, supply chain, logistics, medical, legal, and any other fields involving technical jargon6. The approach enables speech recognition systems to understand industry-specific jargon and vocabulary, enhancing accuracy and responsiveness in complex enterprise settings2.