A major component of this class is the group project you will be working on from now until the end of the semester. Not only is this project a large fraction of your final grade, it also represents an opportunity to explore a topic from class in a more in-depth manner or to try something new and exciting in the realm of AI. The class project will consist of several milestones, including a project proposal (due on 11/4 at 11:59PM), interim project update (due on 11/25), and a final deliverable including a full report and complete code and data (due 12/18).

For this assignment, your group should assemble a one-page project proposal that contains the following components:

Component 1: Introduce Your Group

At the top of your proposal document, list the individual members of your group (names and email addresses). For this project, groups should consist of three or four students. If you are having trouble assembling or finding a group, please let Prof. Scheirer know ASAP. All group members are expected to contribute equally to the project. A group member assessment form will be circulated near the end of the semester to help gauge this expectation.

Component 2: Choose Your Topic

Your proposal's title should reflect the chosen topic. You are free to choose any topic related to AI that interests your group. However, the chosen topic must be of sufficient scope and difficulty to provide enough work for your group from now until the end of the semester. You can also choose one of the following topics, which have already been pre-approved as being sufficiently rigorous:

• Design, implement and test a face recognition system that can identify unconstrained faces in photos.
• Design, implement and test an object detection system that can label objects in photos.
• Implement a semantic segmentation approach that can parse and correctly segment road scenes for autonomous driving.
• Build a system that can generate attribute-based descriptions of people in photos.
• Train a generative model (e.g., a GAN) that can synthesize novel instances of known objects.
• Build a chatbot which engages in naturalistic dialogue (maybe even fooling the class into thinking it's a real person).
• Create a text processing system that can label the sentiment of comments left on Internet discussion boards.
• Design a feature set and classifier that can be applied to the problem of authorship attribution.
• Implement a new AI engine in an existing open source game.
• Create your own video game with an adversarial AI component.
• Apply machine learning to the problem of network intrusion detection; use real tcpdump captures to evaluate your approach.
• Write a speech recognition system that is able to recognize commands uttered by anyone to control the UI of an OS
• Implement a speaker identification system that is able to recognize individuals by the way they talk.
• Design a system that is able to recognize the genre of music that is provided as input.
• Design a new segmentation approach for the connectomics neuron segmentation problem.
• Implement and test a hyperparameter optimization method on multiple problem domaiins.

Component 3: Describe the Elements of Your Project

The bulk of your proposal should describes the elements of your project in sufficient detail. First, outline the specific tasks you hope to accomplish. Assign estimates to how long you think those tasks will take to complete. What language will you use to write your software? List any development tools and software libraries you intend to use. For this project, you are encouraged to use existing open source libraries to support your project development where appropriate. If your project is data-driven (i.e., you plan to use machine learning), explain where your training, validation, and testing data is going to come from. Identify specific criteria for success (how will you know if your project outcomes are favorable?). Make sure you describe how the software will be used in practice (i.e., how do things look from the user's perspective?).

Submitting Your Proposal Document

Each group needs to submit one copy of their proposal (in PDF format) via email to Sophia (sabraha2@nd.edu). Prof. Scheirer will grade these and provide guidance on the difficulty, scope and timing of the overall project. Feedback will be returned to you shortly after the submission deadline.

If you have specific questions about this proposal before it is due, please reach out to Prof. Scheirer or Sophia during office hours, or via slack. This proposal is worth 100 points.