PhD projects in the VIA Group

The following is an overview of what you may expect as a PhD student in the VIA research group. I advise all new students that one of the most important decisions that they need to make is the selection of their thesis advisor; a good working relationship between the student and advisor is very important to the success of a PhD program. The following provides information on my philosophy with respect to PhD students. The main text has been taken from the web pages of my Cornell colleague, Professor Edwin Kan. There have been surprisingly few small modifications to match my personal philosophy.

• Ph.D. Thesis topics:
  The pursuit of Ph.D. research would require the ability to work independently on a potentially nebulously defined topic and might also involve significant stresses associated with possible lack of measurable progress and a number of required exams. The procedure below is a general guide line, instead of rules. Your Ph.D. thesis topic in my group will be developed in the first year after some in-depth mutual understanding based on your course performance, research aptitude and career goals. Ph.D. thesis should focus on an innovative concept, whose proof or realization is the center of mass of your research work. Thus, a straightforward implementation with known methodology is in general not a good candidate for Ph.D. thesis, or anything that can be finished in an extended undergraduate or MEng project is also out of the question. I usually will assign a small trial project in the first year. Due to the nature of image analysis research, your thesis usually contains the components: experiment design including image data-set documentation, algorithm design, and formal algorithm validation. Sometime there is an exception, but the development in all three areas will give you good versatility in the long run whether you work in industry or academia. Switch of thesis topics after the first two years is still possible, but it is usually associated with delay in graduation.
• Expected number of years in the Ph.D. program:
  This is the FAQ (frequently Asked Questions) from my past experience, but I do not have a good answer. From both your and MY points of view, timely graduation is EXCELLENT. However, graduation with no substantial research work from a Ph.D. program is almost suicidal in your professional career. As a rough estimate (not intended for serious auditing, and I say this mainly to protect your benefit), you will need to submit/publish two to three journal papers on your own Ph.D. work before graduation (for sure, some journal paper materials can be published previously in conference papers, which are strongly encouraged). Although the more publications usually the better, quality of the paper is of ultimate concern, since it is your/my permanent research record. What is really important is how we can push your career to a higher point after Ph.D., not just within Ph.D. Finally, the number of years (I am not trying to avoid this question): it will take a true and "lucky" genius to graduate from Ph.D. in less than three years, with the understanding that some of you are true genius. I will be also very unpleasantly surprised if you stay in the Ph.D. program for more than 6 years, which I will give you a lot of help and pre-warning when I sense the lack of progress in research. The shortest I recall is three years after the master's degree. I would say the average is about 4.5 years.
• Collaboration and interaction within the research group:
  Collaboration is highly encouraged within the group, within Cornell, and even with other academic programs. You may work closely with one or more senior graduate students when you start. There should be no barrier among the research work in our own group, and I fully encourage free and open academic practices (for sure, when the research work is not mature, you do not publicly talk about it, not because of protection of "intellectual properties", but because of the relatively high probability that the work may contain serious faults). Please take initiatives instead of waiting collaboration to come at your door. Joint authorship within the group is highly encouraged. Much of the research in image analysis involves collaborations with other groups in which the application for the target of the image analysis resides.
• Fincial support in the Ph.D. program:
  I am committed to financially supporting all my Ph.D. students through research assistantship, fellowship and teaching assistantship (the order represents my preference). However, research funding comes and goes (especially these days). I will try my best, and hopefully you will try your best too. Your thesis topic does not need to tie to a specific contract, and should have more global viewpoints. Another concern is that I need QUALITY teaching assistants for my courses at Cornell to offset my loads, and usually teaching assistantship for one semester to a year is a good experience to most Ph.D. students for testing water of a possible academic career. The teaching assistantship is not a requirement, but maybe a reality. This usually happens in the first two years of your Ph.D. program.
• Course and exam requirements in the Ph.D. program:
  Although there is no official course requirement in the Ph.D. program, I do encourage you to take and excel in Cornell courses. Curriculum design will be different for individuals, but graduate-level courses provide a systematic approach for a topic and will broaden your knowledge. In Cornell EE Ph.D. program, there are three necessary exams: Q exam, A exam and B exam. Graduate committee formation will come from mutual agreement. Q exam is meant for comprehensive knowledge and methodology to approach a research problem, and is usually taken within the first year of your Ph.D. program. The "A exam" is meant for specific knowledge and techniques required for the completion for your Ph.D. thesis, and usually contains a 25-40-min presentation by the candidate on the proposed Ph.D. topic, which usually occurs in the third or fourth year of your graduate study. The B exam is a presentation of your Ph.D. work with totally independent handling of questions, whose successful outcome results in the shaking hands of "doctors..."
• On working and working in the group:
  I personally am very committed to the VIA research program and devote a great deal of my time and effort to it; I also have high expectations for my students. We have group meetings typically held on a weekly basis in which the activities of the group are reviewed and on occasion presentations are made by myself or other group members. Although I encourage graduate students to spend a significant time in office or lab, I do not monitor student's time/place management (I do monitor the progress though). When I am in my office, I am generally available to my students or will find time to meet with them as my commitments permit. Do feel free to drop in my office for discussion any time (Yes, I mean it). My most important job at Cornell is advising Ph.D. students, i.e., you. If I am too busy preparing a proposal or a talk, I will tell you frankly to leave me alone and my next available time (please feel no offense...) I have tried to reply to most of the email inquires from serious applicants, but I do NOT reply to chain or blind emails and I am not able to respond in a timely manner to many emails. If I detect that the email is broadcast to hundreds of professors without sincerity, it actually has an adverse effect. I hope everyone can be considerate for other's time, instead of abusing the convenience provided by Internet.
• Contact my graduate students for a story from the other side:
  In my prejudice, I am surely the most reasonable advisor and the best teacher (yes, by definition)... However, you are welcome to converse with my graduate students about my advising and teaching. Decision on a Ph.D. advisor is as important for you as it is for me. The list of my current graduate students and their email addresses are in my main web page.