As you can imagine we will have a large number of very qualified applicants and limited funding for assistantships. I would encourage you to talk with various faculty and let them know your interests as you have done here.
so list specific names in your statement of purpose. This will help the admissions committee identify who should review your application, or who should take a look after an acceptance decision is made.
I cannot commit to taking on any new students until we go through the process of admissions and make offers. And unless there are extraordinary circumstances, we cannot fund MS students.
If you have not already done so, please send me your CV, grades and a statement of research interests. Once you have applied, please let me know and I will watch for your application.
I would also encourage you to visit the Departments site (https://engineering.buffalo.edu/computer-science-engineering.html) and the AI Institute's Site (https://buffalo.edu/ai).
Good Luck in furthering your education.
The normal process is to apply for the program, and that will expose your material to the entire department.
As you can imagine we will have a large number of very qualified applicants and it is difficult to evaluate all of you before seeing a complete application. I would encourage you to talk with various faculty and let them know your interests as you have done here.
Unless there are extraordinary circumstances, we cannot fund MS students. The amount of time that you're here, and the course load simply do not allow. I do take many independent studies from master students so if you're interested in research that's one way to get exposed to it. Proving yourself during your MS program is a great way to begin working with faculty, and increases your chances of admission to the PhD program.
If you know you eventually want to do a PhD, and are very interested in research, I would encourage you to consider applying to PhD programs directly (not just at UB, but in general).
I would also encourage you to visit the Departments site (https://engineering.buffalo.edu/computer-science-engineering.html) and the AI Institutes Site (https://buffalo.edu/ai).
In the past, we have been approached by many students wishing to complete a master's thesis. Students should first decide if they want to do a master's thesis or perhaps an independent study with a research component. There is a significant difference in the expectations of what a thesis involves, and the quality of the final product.
We will be very selective in the students that we take for master's work. They must have a solid background in computer science with the ability to program and implement various algorithms efficiently and quickly. Students must have a solid work ethic and be willing to devote a tremendous amount of time to make a significant contribution to the field. Unlike independent studies, a master's thesis will be the student's unique work. It is possible to work with other students on a similar topic, but each student's contributions must be unique and significant.
Students who choose to do a master's thesis should expect to start by reading a significant amount of literature so that they can define a substantial problem. This literature review is an essential part of one's thesis. It sets a baseline for the research and allows one to defend their contributions. Something that must be done upfront and cannot wait until the last minute while writing the thesis.
One thing that we will insist on is that students provide a research proposal. This proposal should be in the form of a white paper and should seek to answer the following questions:
After we have agreed on a topic, used to think hard about other details, including:
Before committing to take on a student's master's thesis and be their faculty advisor, we must ensure that you have the ability, not just the background, to do research. In some cases, if the student has prior research background, including publications and a very well defined problem, this may be sufficient. Otherwise, we will require you to do an independent study first. We realize this is a challenge for students that want to stay an extra semester, but we feel it is necessary to enforce.
Finally, the student must be prepared to work on something that is of interest to their faculty advisor. Unlike an independent study where the student fixes the topic and is primarily responsible for what content goes into the end of the project, a master's thesis is a collaboration where the faculty advisor is very involved to the point of multiple meetings a week. We expect publications to come from this work and for there to be a significant contribution.
Currently, our lab is interested primarily in artificial intelligence and machine learning applications of computer vision and, in particular medical analysis and media forensics. Another area is the intersection of language and computer vision. If students are interested in natural language processing, there are options here as well.
Students should consider the content of this document very carefully and do your homework before deciding that you want to do a thesis. Our bandwidth is limited, and we want to give as much research time to students as possible.
Please inquire with additional questions
This document provides a general outline for those students at UB that might be interested in an independent study. In past semesters, there have been challenges with students who are not sure exactly what they want to do, but feel that they want to be exposed to research. We want to help as many students as we can. However, we commit significant time to various projects, and we want to make sure that we can do the same for independent study students, and the relationship will be mutually beneficial.
This document intends to make it very clear what is expected of an independent study (Undergraduate or Graduate), should we decide to work together, and the way that the course is graded.
One important thing to remember is that an independent study is just that, an independent study. It is something that students want to study or research, that is important for their future, and that will benefit from the help of a faculty advisor. It is not a seminar where students will learn primarily from lectures given by students and faculty.
The first thing that a student needs to decide however is if they want to study a subject in detail or do research. A study, for example, will allow one to take a topic, become intimately familiar with it, then take the topic and apply it to a new or existing problem, dataset, or topic. However, by and large, students will not make a broader attempt to solve the problem, only report on the results of the application of what he or she has learned. A study is a good approach for something that will be limited to one semester. The study can lead to one's Master's paper, but typically NOT a thesis.
Research, however, will require defining a problem first, then taking a broad approach to solving it. The research could lead to a master's thesis if the student decides to do one, and it may be challenging to complete a project in a single semester.
If a student chooses the study, they will work on a specific subject, understand it in detail, and apply it interestingly. If a student chooses the research, they will need to understand the subject well enough that they can contribute to the topic that warrants publication. We can provide more details on each moving forward
Students will be expected to produce a RESEARCH PAPER at the end that is suitable for publication (in a workshop or conference if you choose research). In that sense, there is a research component. It is not necessary to publish it, but that paper will include everything from a literature review, to a description of the problem, to the approach, to the data used (i.e., the data and evaluation protocol) and a summary of your results. It will not be possible to get an A in the class without producing this report by the end of the semester.
Students will be graded on the level of effort they put in during the semester and on the final project. Students will all have read or should have read technical papers. Think of the final report as a "conference paper length" (7-10 pages) contribution, which fully describes your understanding of the problem or domain, how you applied it, and the results you obtained.
While we have some projects that we are interested in having students work on, it is also possible to come with one's project. When one does, they should understand the nature of the problem, and have some idea of how they are going to attack it.
Students should not come just say that "I want to do machine learning" or "AI" unless they are interested in one of the projects that I am offering.
A good topic will be one that has not been actively addressed, a problem that needs a new approach or a problem that has been addressed, but that one can hope they can make a significant contribution too. Often one of the challenges with any research problem is data. We encourage students, therefore, to pick a topic that has publicly available data for evaluation, and if possible, and challenge already set up so they can measure their progress. Various conferences have these, but there are also new datasets and problems on Kaggle, for example.
If students have not done research in the past, and if they do not have a well-defined project, we would prefer it if there are other students to work with to make a more significant contribution to a project. If, however, students do have a research background and are confident that they can make produce a quality technical report, we will be happy to consider a proposal.
What I would prefer to do is have a meeting of students who are interested in doing an independent study right at the beginning of the semester. Or possibly a couple of weeks before.
If students are still interested, please let us know, and we will schedule a meeting.
Also, please send us a current resume and statement of interest (if you have not already done so) so that I can have some context when we talk.
Unless you are self funded, we do not have funding for short visits. If you are totally self funding, it may be possible to have you work with one of our faculty, but we need to make that decision on a per faculty basis.
Let me know if you are still interested and we can continue the conversation.
