Going all out to race to the top
He Tianrui is one of the outstanding graduates of the Department of Mathematics. Except for majoring in Mathematics, he also chose to minor in Computer Science. He shared with us about the reason he chose to study Computer Science, ‘Mathematics is my own interest. I don’t regret although it is a difficult subject. For Computer Science, it is because it can apply the knowledge into the industry. The knowledge about artificial intelligence I learned from Computer Science could combine with Mathematics, that is also the trend today.’ When talking about the reason that he chose UM, he says, ‘When I was studying in high school in Fuzhou, our teacher once recommended UM to us. I was the only person kept in mind while other classmates were not listened.’ Today, he successfully becomes one of the excellent student of the Department of Mathematics. In regards of his impression towards, ‘UM is excellent in all aspects and it has many learning resources. I was able to join a 3 months exchange program to go to the University of California, Berkeley.’ he says. In addition, the teaching in UM is in English, Tianrui shared that his English level has improved a lot in the past 4 years, now he is able to watch English dramas without subtitles. The improvement of his English level makes him more confidence, even when he meet foreign teachers or go abroad for exchange program.
A speech emotion recognition system help to improve customer services
Tianrui’s final year project is about a speech emotion recognition system (SER system). Through the emotion of the speech, the speaker’s emotion (such as tone) are inputted into the model he designed. This speech emotion recognition system can be applied into a company’s customer service. If guests contact the customer service, the system can judge customer’s emotion. Whether the call is about complaint or thank you, the system can be transferred to the corresponding customer service staff. The idea of doing this topic is from his full time internship at Huawei. He saw many customers contact customer service for different questions. If they can rely on the recognition system to identify their needs, it will greatly improve working efficiency and customer satisfaction. Therefore, he believed that there are markets for this system and worth to be further investigated. The successful rate of the model designed by Tianrui is about 65%. Since he confronted many challenges in the whole research process, ‘this project not only improved his programming skills, but also developed his problem solving skills, he also learned how to adopt previous research results in his own research and wishes to continue to optimize the system in the future.’ he says.
Successfully accepted by numerous well-known universities
Tianrui, who aims to continue his study journey, was successfully accepted by several well-known universities for his master degree application, including the University of Hong Kong, the Chinese University of Hong Kong, the Hong Kong University of Science and Technology, the Australian National University and the University of Melbourne. After being accepted by many prestigious universities, he wishes to thank his supervisor Prof. Liu Zhi. Except for guiding him to complete the project, Prof. Liu Zhi, who completed his Ph.D. degree in Hong Kong, have given him many valuable advices, which made him finally decided to go to Hong Kong. He also encouraged his underclassmen, ‘If they want to apply for a master degree then they need to study hard to increase GPA, and better equip themselves to improve the English level.’ Tianrui finally chose to study at the Chinese University of Hong Kong, he also wishes to fulfil his immense potential in a new university and environment in the coming year.
天睿的畢業設計項目是有關語音感情識別系統(SER system)，透過說話內容，把對方的感情 (如聲調語氣或關鍵字) 輸入到自己設計的模型中，從而判斷對方的情緒。這個語音感情識別系統可以應用在企業的電話客服上，若客人打電話到客服來，系統就可以判斷對方的情緒，無論是有關投訴或致謝，都可以轉接到相應的客服人員。畢業設計項目的靈感來自於自己大四在華為實習時，當時天睿見到很多客人因為各種問題打電話到客服，若能依靠系統成功分辨並分流客人，將大大提升客人滿意度。因此，他認為有一定市場及條件去研究。這次設計的模型成功率約為65%，天睿說:「這次畢業設計項目的模型是要用代碼寫出來，而且當中牽涉到各種難點及錯誤需要深入分析，不但提升了自己寫代碼的能力，同時也提高了自己找出及解決問題的能力，也學習到如何把前人的研究成果用在自己的研究上，期望未來可以繼續優化這個系統。