aiOla’s proprietary model significantly improves on OpenAI’s Whisper, achieving a 45% increase in speech recognition accuracy when transcribing domain-specific dialogue
aiOla, a leader in speech recognition technology, has announced a new AI model that leverages their breakthrough research into jargon detection, allowing it to instantly adapt to the unique vocabulary of any industry without the need for re-training. aiOla is enabling enterprises to effectively capture valuable, previously uncaptured data by utilizing speech recognition technology specifically tailored to their business jargon and needs. aiOla’s technology replaces manual processes previously done with pen and paper, while supporting over 100 languages and accurately transcribing even heavily accented speech.
Off-the-shelf speech recognition models, including leading solutions like OpenAI’s Whisper, fall short in industry use due to their inability to accurately transcribe domain-specific terminology. To address these limitations, extensive training tailored to each industry’s unique requirements is usually necessary. The initial cost of training state-of-the-art AI models can reach hundreds of millions of dollars, and even the process of fine-tuning models is extremely resource-intensive, requiring specialized AI expertise.
aiOla’s model, leveraging their proprietary technology, provides the flexibility enterprises need across all industry sectors, including manufacturing, supply chain, and beyond. Through an innovative model architecture that leverages prompt guidance, it effectively incorporates domain-specific jargon, enabling bespoke AI speech recognition systems with zero retraining.
aiOla has begun deploying this technology in Fortune 500 companies in areas such as logistics, shipping, manufacturing, maintenance, and inventory control, offering services tailored to companies of any size while producing immediate, measurable ROI.
“Enterprises across every industry are acutely aware of the pressing need to adopt AI to maintain a competitive edge, but they don’t know where to begin,” said Mitch Garber, Executive Chairman of aiOla. “While text-based AI solutions are great for office environments, speech interfaces reign supreme for industrial settings because they seamlessly integrate into existing workflows and collect previously uncaptured spoken data. Prior AI speech recognition models couldn’t perform for business use cases because of their inability to grasp jargon. Today, aiOla is changing that by providing instantly tailored AI models that can understand the unique jargon of your specific industry, your organization, or even your team.”
Also Read: UiPath Announces New Platform Features to Enable Organizations to Build Better, Faster, and More Comprehensive Automations with UiPath Autopilot™ and GenAI
aiOla has published research outlining their novel approach to achieving instantaneous jargon recognition. They use a two-step process: initially, the presence of specific terms is detected through aiOla’s advanced keyword spotting model, and this information is then utilized by their Whisper-based model to augment its overall speech recognition capabilities and correctly detect the jargon words or terms. “For this use case, we decided to enhance the most accurate speech recognition model out there, OpenAI’s Whisper,” said Gil Hetz, VP of Research at aiOla. “However, you can apply this approach to any speech recognition model, including Meta’s MMS model and proprietary models, unlocking further potential to elevate even the highest-performing speech-to-text models.”
aiOla‘s model’s capability to understand jargon instantaneously is achieved by initially freezing the main speech recognition model and adding a proprietary adaptive layer. This adapter undergoes one-time training to guide it in effectively utilizing a vocabulary of jargon, while the model’s core capabilities of general-purpose speech recognition are retained. After training, the jargon vocabulary can be hot-swapped to the jargon of a different sector, achieving state-of-the-art performance in recognizing both industry-specific language and general speech.
Source: PRNewswire
