Since the debut of the original Star Trek series, people have always dreamed of owning a pocket-sized device that can translate languages in real time. The challenge of deciphering a doctor's handwriting has also long been a common frustration, fueling the demand for more efficient and accurate language solutions.
Thanks to breakthroughs in deep learning and AI technology, Keda Xunfei, one of China’s leading AI companies, has developed practical and innovative solutions. By leveraging well-trained neural networks, they are able to apply their knowledge effectively, bringing advanced translation and voice recognition capabilities to everyday users.
"Speak to Me" is a portable, handheld device designed by the University of Science and Technology. It translates spoken language into another with remarkable accuracy. Powered by NVIDIA Tesla P4 and P40 accelerators and Keda Xunfei’s cutting-edge translation algorithms, this compact device supports real-time translation between Chinese, English, and Uighur, making communication across languages easier than ever before.
As the largest provider of voice technology in China, HKUST has made significant strides in this field. Zhang Zhijiang, associate dean of the Cloud Computing Research Institute, described the device as “a professional translator in your pocket,†helping users overcome language barriers anytime and anywhere.
One of the key advantages of this system is its impressive 15x improvement in deep learning efficiency. According to Zhang Zhijiang, the speech recognition accuracy on the Xunfei Voice Cloud platform reached an impressive 97%, significantly outperforming traditional CPU-based systems. These advancements have opened up new possibilities for AI-driven applications at HKUST.
Since its launch in March, Keda Xunfei’s portable translation devices have been sold in over 97 countries, delivering more than 2 million services globally. This success highlights the growing demand for real-time language solutions in today’s interconnected world.
Beyond translation, HKUST is also focusing on smart healthcare applications. In collaboration with numerous hospitals and healthcare organizations, the company is integrating its voice technology and AI expertise to enhance medical care for both doctors and patients.
No longer do doctors need to spend hours trying to read their own handwritten notes or prescriptions. With HKUST’s technology, doctors can simply speak into a microphone, and the system automatically converts the audio into a complete electronic medical record, streamlining workflows and improving patient care.
HKUST’s voice cloud platform can convert even the most difficult-to-read doctor’s notes into clear, readable text. Leveraging the power of the Tesla P4 GPU, the intelligent dictation application quickly filters out irrelevant information and transforms medical data into structured text. The system achieves an impressive 97% accuracy rate in Chinese language conversion.
Keda Xunfei uses its voice cloud platform to process massive volumes of data in the cloud, offering a range of intelligent voice services such as Xunfei Input Method, Xunfei Language Point, and Consonant Voice Assistant. With 1.5 billion users and an average daily service volume of 4 billion, the platform continues to grow in popularity and impact.
“As the Xunfei Voice Cloud continues to evolve and expand its applications, NVIDIA’s deep learning GPU platform provides the necessary technical support for Keda Xunfei. This enables users to benefit from the latest technologies and enjoy a better overall experience,†said Zhang Zhijiang.
Ultra Thin Switch Power Supply
Switch Power Supply,Switch Mode Power Supply,Switch SMPS LED Power Supply,Ultra Thin Switching Power Supply
FOSHAN SHUNDE KELICHENG POWER SUPPLY TECHNOLOGY CO., LTD , https://www.kelicpower.com