FAQs

The main objective of this course is to offer professional Masters students a combination of industry mentorship and hands-on data science training. It aims to provide students with practical experience on significant projects beyond the typical classroom setting.

 Students collaborate in teams of 3-4 individuals. These teams work on innovative research projects that are proposed by industry partners.

Each team will receive guidance from an industry mentor, an additional Ph.D. student mentor, and the course faculty instructor.

Collaborating with industry partners provides numerous advantages. These include access to rich, industry-scale data, insights into real-world challenges, and the opportunity to forge valuable industry connections that can be beneficial for future endeavors.

To apply for the course please complete this form. Please fill out the form before Friday, November 3, 2023.

In order to apply, the student should have completed or be on the path to completing two data science core courses with a grade point average of 3.0 or higher. Apart from the data science core courses, we will also consider the following courses as satisfying the course requirements:

646 (Information Retrieval), 683 (AI), 688 (Probabilistic Graphical Models), 674 (Visual Computing), HONORS 449C/DM (Honors Thesis in Machine Learning)

Every student who applied will receive an email with the results of the selection process around mid-November.

This is an extremely popular course. This year, we received more than 220 applications from highly qualified students. We believe that with the right guidance, all students can succeed in the course. However, with only 80 available seats in the course, we try to select students who are best positioned to take advantage of the project-based learning opportunity that the course provides. We use holistic evaluation criteria for the selection process. The two most important aspects of the criteria are the following:

1. Coursework: We require a minimum of two relevant graduate-level data science courses. Most students who are selected often have three or more relevant courses completed. The minimum course requirement can be satisfied by two data science core courses or using the list of important courses given below.  Note that we give a higher score to 600-level courses.

List of important courses:

• Reinforcement Learning (COMPSCI 687)

• Advanced Natural Language Processing (COMPSCI 685 or 690N or 690D)

• Optimization (COMPSCI 690OP or 651)

• Computer Vision (COMPSCI 670)

• Neural Networks (COMPSCI 682 or COMPSCI 691NR)

• Machine Learning (COMPSCI 689)

• Visual Analytics (COMPSCI 690V)

• Intelligent Visual Computing (COMPSCI 674 or 690IV)

• Algorithms for Data Science (COMPSCI 514)

• Information Retrieval (COMPSCI 646)

2. Projects and relevant experience: We encourage students to talk about their relevant experience in the application in detail. While course projects count as relevant experience, most students who are selected have additional experience working on research projects outside the classes either through other independent study projects, through independent research projects, or through industry experience. In the description of the relevant experience, we are looking for several key characteristics that may increase a student's chances of being successful in the course. Following are some examples: 

   1. Whether the student has demonstrated that they take initiative when required

   2. Whether they can step up to take complete ownership of a project if required

   3. Whether they have demonstrated all or most of the skills that may be required during the course. A strong project experience will demonstrate contribution to several critical steps of a project's journey like identifying a problem, identifying suitable data and modeling approach through a literature survey,  and presenting the results to the stakeholders through a paper, report, or presentation.

Note that we don't necessarily care about the outcome of the projects in terms of impact, but we care more about what relevant skills you may have learned through your previous projects. 

The process is structured to ensure that students are matched with projects that align with their preferences. Here’s a simplified overview:

Collecting Project Preferences

• The students who are enrolled in the course receive a project preference form and project descriptions. Students indicate their preferences by scoring each project on a scale of 1 (low preference) to 5 (high preference), and students are encouraged to provide scores for as many projects as possible to optimize matching.

Matching

• An algorithm called fair matching uses students' preference scores for an initial automated matching process. The algorithm tries to satisfy as many students as possible.

• Some manual adjustments may be applied to the initial algorithmic assignments to ensure appropriateness. 

• The complexity of the matching process means that accommodating special requests may not always be feasible. 

• Students will be informed about their respective project assignments a week before the course commences. 

Students who are selected to enroll in the course will receive the project preference form around mid-December first week of January.

During the weekly class sessions, students will receive professional development education, training in data science software infrastructure, presentations on data science research, and career guidance.