Keda Xunfei held the "Artificial Intelligence + Total Innovation World - 2016 Science and Technology News Annual Conference" at the Beijing National Convention Center. The conference showcased the advancement of AIUI, the user interface of HKUST's artificial intelligence, in terms of products and markets. AIUI is a human-computer interaction product developed by Keda Xunfei based on speech recognition and combining image, visual and gesture recognition. Its integration includes dual-worker technology, microphone array technology, voiceprint recognition technology, dialect recognition, semantic understanding technology and And content services, and can be combined with many artificial intelligence products, so that artificial intelligence products can operate under the interface of AIUI. Since its launch in 2015, AIUI has been used in human-computer interaction applications in education, government, and transportation. At present, the data given by the University of Science and Technology News shows that the number of users of AIUI has reached 890 million, and the average daily usage is 3 billion times. In the future, human-computer interaction is still the focus of Keda Xunfei on speech recognition technology. The main products and market conditions introduced at the conference are as follows: Voice dictation AIUI's voice dictation function has been able to directly modify, translate, and punctuate with voice. At the same time, it can remember names, response speed is less than 200 milliseconds, error rate is as low as 2.24% in noisy environment, and can be more accurately identified. Dialects and accents. In addition, the speech recognition system also has an encryption function for the confidential scene. Currently, AIUI's speech recognition is mainly used for instant submission of meeting minutes. This year, the speech recognition system also cooperated with the Supreme Law, mainly for trial records. According to data from the University of Science and Technology, the time for the highest law to handle the trial records of ordinary cases has been accelerated from the original three hours to five minutes. 2. Real-time translation The real-time translation system of HKUST can be used on the website or download the app. It can recognize the voice, translate it in real time and put it out. It is mainly used to communicate with people of different languages. At present, the Sino-Vietnamed real-time translation system has emerged and played a role in the work of aiding Xinjiang. It is used for exchanges between the aid workers in Xinjiang and the ethnic minorities in Xinjiang. Keda Xunfei said that the company is currently in the world leading level in the field of Chinese-English translation. In the future, it will develop a real-time speech translation system in the world's major languages ​​to solve the obstacles in people's communication in different languages. 3. Human-computer interaction in smart home The cooperation between Keda Xunfei and smart home is currently concentrated in the TV field, and the product is Xunfei TV Assistant. The University of Science and Technology will implant its own voice interaction system into the remote control. As long as the user speaks the command to the remote control, the TV will respond accordingly, such as screening the movie and adjusting the volume. The product is now able to support deeper human-computer interactions, such as some more casual instructions, such as "the end of the music is very good, what is this song", the system will display the song's title. At present, HKUST has cooperated with more than 3,000 manufacturers in the field of smart homes, and has launched more than 1,000 products this year. In the future, the voice interactive system of HKUST will also be embedded in more smart homes, and people can use voice to control more home appliances, not just TV. 4. Car voice interactive system The latest product of the Keda News Flying Vehicle Voice Recognition System is the Flying Fish Assistant, which can complete navigation, find restaurants, play music, and make calls by voice dialogue. This is related to AIUI's speech recognition supporting speech modification and learning names. Even when you are on the phone, the system finds the wrong address book friend based on the voice, and can also modify it directly by voice. When using the flying fish assistant to navigate, the system can identify the short name of the place name, alias, and so on. In the future, HKUST will hope to make progress in the direction of autonomous driving, hoping to make the autonomous driving environment more humane. This requires more human-computer interaction, and the machine can engage in deeper dialogue with humans. 5. Financial services robot The “Xiaoman Robot†launched by HKUST has been used in banking services. The financial services robot recognizes the user's voice commands and pushes relevant financial services on the screen. In addition, "Xiaoman Robot" can also guide the user in the bank lobby and take the user to the corresponding window to handle the business. 6 personalized education The automatic composition review system developed by Keda Xunfei has been applied to the scoring of the 4th and 6th grade exams, which can score the test papers and propose amendments. In addition, Keda Xunfei's digital personalized recommendation education system can analyze the error rate of each topic in the exam, and list typical errors. According to the different error conditions of different students, push different misunderstandings and provide personalized learning content. And save the teacher's unified explanation time. At present, the system has been applied to the education of schools such as Beijing Normal University Second Middle School and Huanggang Middle School Guangzhou Branch. In December 2014, HKUST announced the “Xunfei Super Brain†program, with the goal of allowing robots to go to college within 3 to 5 years. In addition to the above functions, the purpose of the artificial intelligence technology promoted by the University of Science and Technology is to learn machine learning. By summarizing the knowledge and errors of all users, the machine continuously recognizes, learns, and iterates, achieving higher learning efficiency than humans. As for the ultimate goal of "three to five years of college entrance examination", there is still no conclusion. 7. Speech imitation "Xunfei Duoying Pavilion" is one of the main products of Keda Xunfei in the field of voice imitation. At present, the mobile app is open for download. This product can mimic the sounds of different people, including tonal sounds and tonal habits when speaking. The voice imitation system of Keda Xunfei mainly uses the direction of film dubbing, which saves the actor's post-dubbing work, and directly allows the machine to replace the actor to complete the dubbing work. In addition, ordinary users can also use the voice imitation system to set the sound of the smart device, such as setting the sound of the storytelling of the mobile phone as a parent, setting the voice of the other party between the couples for navigation, etc. Ptfe Filler Eptfe Filler,Spiral Wound Gasket,Eptfe Tape For Swg,Expandable Ptfe Filler Gaskets Tape Cixi Congfeng Fluorine Plastic Co.,Ltd , https://www.cfptfeseal.com