Let me make a brief self-introduction.
My name is Liu Jingyong, and you can call me Jason. I’m studying for master degree in Tongji university, majoring in Computer Science.
At the first two years of my university life, I devoted myself to the major courses and got good marks. It might be a proof that I’m awarded the first-rate Scholarship twice. When I was in the third year, I realized that quite a few companies emphasize on professional skills and working experience, therefore I put more time to improve the technique skills, such as C, CPP, Visual CPP and Java. Meanwhile I began to hunter for some relative part time jobs.
In the end of 2002, I was recommended to study for master degree in CS Dept. without taking the entrance test. Since then, I became a member of Tongji Artificial Intelligence Lab. There, much of research work helped me cultivate the capability of analysis and problem solving. Also, I learned lots of communication skills from my colleagues, who are always willing to share their wisdom and ardor with me.
My research work in our lab includes three parts:
The first is auto-segmentation system for natural speech. We cooperate with Motorola China Research Center to finish the project. Simply speaking, the system is utilized to divide the natural speech into small monophone units.
The second is Text-To-Speech system of Mandarine. The project occupied us nearly one year and a half. Like its name, this system can transform Chinese text to continuous speech.
The third is the research of digital watermark of audio signals. My responsibility is to research on how to embed watermark into digital audio. The research may make great sense in the copyright protection issue.
And then I would like to talk something about my technical skills.
I’m quite familiar with C/CPP/Visual CPP, and have rich experience of programming with MFC, STL. Java is also one of my programming languages. Eclipse and JBuilder are my favorite Java IDE tools. I am also able to use Linux platform. Last year I acted as a support professional in Wicresoft Shanghai for 3 months, there I gained deep understandings of Windows platform. Some database development tools, such as SQL Server, PowerBuilder are also available to me.
My biggest merit is the self-study ability and the independency when tackling tough problems.
As we know, in IT field, technologies update very quickly. It is difficult and unnecessary for us to master all of them. So the ability to study rapidly is very important when we need to use new technology which we never touch before. I think all along I am making efforts to build up this kind of qualification. For example, once I was required to complete some modules in a Java project, but at that time I had little knowledge about it. The only way for me is to master Java as soon as possible and put what I’ve learned to practice. Within a short period, I got the key points of Java and accomplished the task assigned to me on time.
When it comes to independency, it doesn’t mean I would not like to communicate with others, but means I’m good at tackling problems independently. For example, when I am confronted with a tough technique issue, first I will make use of all the related tools at hand, such as Google or my books, to resolve it. I will not seek others’ help at the very start without any effort to resolve it by myself.
As to the sense of responsibility, I attempt to form the habit of taking all aspects of a piece of work into consideration, analyze the advantages and disadvantages, secure that no problem would occur and responsible for what I need to complete. It is my target to keep everything going as planned and I will never give up easily.
My demerit is that sometimes I’m a little stubborn. For example, when I’m working on a task and encounter some difficulties, maybe it is due to the wrong method that I’m using. However, I’m not inclined to give it up, because I think it is very pitiful that I have to do it again from the beginning regardless of so much work I have done. That is not a good habit. Actually, we should learn to give up or adjust the direction of the work at a proper time, and reevaluate the targets that are accessible.
I have encountered a big difficulty when implementing the TTS system. Our system is based on a large corpus. That is to say, to a Chinese character, we have plenty of choices in the database and we must find a most suitable one so that we can use it to synthesize the speech. If every time we search it from the whole list made up of the Chinese character, the efficiency of the TTS system will be unbearable. Therefore, it is urgent for me to find a good method to divide every high-frequency Chinese character into several groups. But how to divide is a big problem. I tried quite a few methods and spent a great deal of time on it. Finally, I found an algorithm called CART (Classification and Regression Trees) could meet our needs. After I introduced the CART algorithm to our TTS system, the efficiency issue was resolved. That’s the biggest difficulty I have met with during my research work.
My biggest failure is that I had to take the National College Entrance Examination twice. The first time in 1998, I got 628 points and applied for the computer science major of Nankai university. Although my marks achieved the basic requirement of the university, I was refused by the CS major. Since I preferred the computer science very much, I didn’t accept being transferred to other majors. So, I returned to my high school with a broken heart. At that time I was really very frustrated and felt the life was so tragic. To comfort me, my parents bought a computer for me and encouraged me to confront the reality bravely. That’s my first computer in my life. In 1999, I participated in the Entrance Examination for the second time. This time I got 644 and chose Tongji university, computer science major. I was successful to achieve my goal. Three and half years after that, when I knew I had been recommended to further study for master degree, I couldn’t help crying when I recalled this unforgettable experience.
I’m interested in the hardware information. Pchome and Pconline are my favorite websites. In my spare time, I pay much attention to the hardware’s development. Sometimes, the outcome of a new technology will make me excited for several days.
I also like writing articles. When I was in high school my Chinese teacher often said that I have some talent in composition. Actually I have great interest in reading and writing articles. The university life gives me more space to develop my hobby. Especially, BBS was a great platform to express myself. When I have some special feelings, I would like to write it down and paste on BBS, share my happiness and sadness with my friends.
Career plan
Computer science is my major and I have determined to work in IT field for my whole life. Just like study, career plan should be accomplished step by step. In the first few years, I’m willing to work from the very bottom, for example, act as a coding programmer. Since technologies update very quickly, I will keep learning new skills without end. My study ability can guarantee my performance in the future. As time going on, I will command some advanced skills and become an experienced employee. To be a senior specialist is my dream at this moment. I believe then I can contribute more to my company, maybe it is IBM if lucky enough to pass the interview. A few years later, after I have made a solid foundation in various skills, including technical and communication, I will put more concentration on project management. At that time, hope that I am able to take charge of large software projects. Of course, that’s a much higher level and needs rich experience and deep foundation.
Why do you choose us?
Your company is a famous one in the industry and boasts a high reputation. IBM’s products, both hardware and software, could be found all over the world. To a student majoring in computer science, it is fantastic to join such a company as IBM.
In my spare time, I pay much attention to information about the company. Up to now, I know that IBM provides plenty of valuable training opportunities to employees. Also, big companies like IBM often have their own culture. Good company culture can benefit me greatly. What’s more, in a foreign company, I have a lot of chances to improve my English, both in oral and written.
Additionally, the post, IT specialist, is in accordance with my background and career plan. As a student of computer science, I have achieved some success in university, especially in academic aspect. Also, I have some related intern and research experience. This makes me suitable for the post.
As a result, I applied for the internship of IBM and choose your company as the beginning of my career. I hope I will be successful to achieve my goal.
What do you want most from your work?
I hope to get two things from my work. The first is professional working skills. Although I have studied in schools for nearly twenty years, I have only learned knowledge and some basic moral disciplines. There is little chance for me to touch the real society and I don’t get enough working practice. So from my work I want to learn professional working skills as much as possible. The second is making a lot of friends. XX owns the most outstanding talents. There are a lot of distinguished people in your company. Making friends with these people will be a great pleasure to me. I hope to share my happiness both in working and daily life with them. Also, I am willing to learn from them and share my knowledge with them. I am sure good friend can help each other develop themselves altogether.
Do you know the saying XXXX? Everyone in Supervising Committee dedicated himself for this saying. To serve the students with our best efforts, I think that’s my biggest achievement during my university life. Our work process can be described as follow: Every afternoon I send out some members to every dining room. They have a fixed position there. When students have some suggestions or questions about the food provision, they will write them down and deliver the suggestions quickly to the manager of the dining room. Then the manager will collect all useful suggestions and take some measures to improve the service quality. In this way, the service of our dining room becomes better and better. In one sentence, we act as a bridge between students and Logistics Group. As the director of this student organization, I feel very proud when hearing friends from other universities say the price is very low and the flavor is very good to have dinner in Tongji’s dining room.
Some students don’t understand our work and they always complain this dish is too sweet, that dish is too peppery and another dish is too salty. We have to explain to him that our students come from all over the country and they enjoy different flavor. We have to satisfy most of them but not part of them. Sometimes our members are often impatient toward these kinds of complaints. I tell them, in any time, under any situation, we mustn’t be impatient in front of our students. I request them to have a smile face when listening to others’ complaints. That’s a real difficult thing for our members who have never received professional training.
The first project is auto-segmentation system of natural speech. It is based on HMM (Hidden Markov Model). We cooperated with Motorola China Research Center to complete the project. The system is utilized to split the natural speech into monophone units (single Chinese character). For example, the natural speech “我是中国人” can be divided into ‘我’ ‘是’ ‘中’ ‘国’ ‘人’ five little wave files automatically by this system.
In order to realize such a system, first we must build appropriate HM models. My responsibility can be described as follows: First, encode and compress the original wave data. Since there is a great deal of redundancy in the original wave data, we need to encode and compress it. This is one of my duties. Second, evaluate the segment result. In order to evaluate the segment result, we need to synthesize the final little waves. In this way, we can perceive whether our work achieve Motorola’s requirement.
The second project is TTS system of Mandarine. TTS means Text to Speech. Like its name, this system can transform Chinese text to continuous speech. For example, when you input Chinese text “欢迎使用” in the text field, a continuous speech will be synthesized and a voice will read the sentence for you.
The system can be divided into two parts. The first part is text analysis system. It is used to analyze the input text and do some pre-process work. We can get most of attributes of a Chinese character in the sentence through the text analysis system. Since this is a project which has been finished by AI lab two years ago, we can introduce it to our TTS directly. Therefore, we put our main concentration on the second part. The second part is responsible for selecting the most suitable units and synthesizing the speech.
My main responsibility includes two parts: the first is to design the database with SQL Server. Because every Chinese character in a sentence has a series of attributes, we need to store these attributes and prepare for unit selection according to these attributes. I designed the structure of the database and completed the module which can input the results of text analysis into the database automatically. My second contribution is to implement the module of unit pre-selection using CART. Our system is based on a large corpus. That is to say, to a Chinese character, we have plenty of choices in the database and we must find a most suitable one, so that we can use it to synthesize the speech. If every time we search it from all the same Chinese characters, the efficiency of the TTS system will be unbearable. Therefore, it is urgent for me to find a good method to divide high-frequency Chinese character into several groups. But how to divide is a big problem. I tried quite a few methods and spent a great deal of time on it. Finally, I found an algorithm called CART (Classification and Regression Trees) could meet our needs. After I introduced the CART algorithm to our TTS system, the efficiency issue was resolved. That’s the biggest difficulty I have met with during the research of TTS.
The third project is the research of digital watermarking of audio signals. The project is supported by National Natural Science Fund My responsibility is to research on how to embed watermark into digital audio. The research is used to the copyright protection issue. Take mp3 for an example, I can embed some information of the singer into the mp3. We call the embedded information watermark. Of course, people cannot perceive the existence of watermark when they listen to the mp3. When some people copy and spread the mp3 without permission, they cannot get rid of the watermark. Even if they erase the watermark by force, the mp3 will be destroyed and cannot be played normally. So, we can extract the watermark from an illegal copied mp3 and provide some proofs in the copyright issue. The project is also the research area of my master dissertation. Our algorithm is based on spread spectrum and wavelet. It also concerns some knowledge of DSP (digital signal process).
本文来自:3COME文档频道(www.reader8.cn)原文链接:http://www.reader8.cn/data/2007/0106/article_40756.html