Language :
(BRONCI Vice President Kent Chen)
Speech recognition technology is transforming various industries, becoming a crucial tool for reducing labor costs and boosting efficiency. BRONCI has developed a locally tailored speech AI recognition system in Taiwan, which supports input and output across Mandarin, Taiwanese, English, and Hakka. This technology serves as a driving force in digital transformation for businesses, with applications spanning finance, telecommunications, healthcare, law, education, media, and more. Achieving over 95% recognition accuracy, it also fine-tunes language models with industry-specific terms and integrates with large language models (LLMs) to further improve user experience.
BRONCI's speech AI recognition system focuses on local language development, utilizing a single model for seamless multilingual mixed recognition without the need to switch models. This capability is suited for various contexts. For example, in financial customer service, the system accurately transcribes Taiwan's common languages into text, allowing for subsequent semantic analysis and customer sentiment detection for more personalized service. It also integrates with customer service systems through RESTful API, WebSocket, and MRCPv2, enhancing operational convenience and efficiency.
In the public sector, BRONCI's technology is employed for real-time transcription in government meetings, pandemic press conferences, and election debates, achieving smooth multilingual support. Using the 'Taiwan Hakka Speech Database' from the Hakka Affairs Council, BRONCI developed a speech API and collaborated with the Council team to create virtual anchors for the xi ian kiong and Haikou Hakka accents, where broadcast can be conducted by simply entering a script, supporting the promotion of Hakka culture and showcasing the societal value of speech AI recognition.
Most of BRONCI's systems are custom-developed to optimize professional terminology dictionaries based on industry needs, ensuring precise recognition of specialized terms, especially important in fields such as finance, healthcare, and law. Furthermore, the system can adapt recognition speed to cater to different audiences, like younger or older individuals, further enhancing recognition accuracy and user experience.
In the media industry, BRONCI's technology is utilized for television programs on Delay Live, significantly reducing the workload and stress of traditional transcription. By refining professional vocabulary, the system can precisely recognize industry terminology even with varied and dynamic topics, cutting manual labor by 50% and increasing break time for transcription staff by 40%. Staff only need to focus on correcting AI outputs, greatly improving efficiency.
To meet the high computational demands of real-time speech AI recognition, BRONCI has chosen Leadtek's high-performance AI hardware, particularly the NVIDIA RTX 6000 Ada Generation GPU paired with the WinFast GS4850T server. This powerful computing platform delivers stable and efficient support for the speech AI recognition system, ensuring rapid response in practical applications.
BRONCI Vice President Kent Chen noted, “Leadtek's advantage lies in its excellent performance, cost-effectiveness, and professional consulting services. Our clients often require customized solutions, and Leadtek provides precise setup recommendations based on data size and customer needs.”
Amid rapid technological development, BRONCI aims to continuously refine recognition accuracy and expand language support. Vice President Kent Chen said, “We plan to add Hakka dialects such as Dapu, Raoping, and Zhaoan. Once we have a solid foundation in Taiwan's language systems, international speech recognition will also be a future goal. To better serve the growing demand in various industries, we are working on integrating Japanese and Vietnamese speech recognition into our platform, enabling more seamless cross-cultural communication.”
Combining high-performance hardware with advanced technology, speech AI recognition is gradually transforming our way of life. Vice President Kent Chen shared, “Speech AI recognition is the future. As technology progresses, seamless human-machine interaction will soon be realized through accurate digitized speech.” As the technology matures, speech AI recognition will unlock even more innovation and potential.
(BRONCI Vice President Kent Chen)
Future
