Skip to content

CS479: Machine Learning for 3D Data

Minhyuk Sung, KAIST, Fall 2023

Project Mockup

Due: October 9 (Monday), 23:59 KST
What to submit: A write-up (up to three A4 pages; shorter is preferred). No template provided. GitHub repository link.
Where to submit: OpenReview

For each team, submit a write-up including the following:

  • Project title
  • Project track
  • Submission ID (assigned by OpenReview)
  • All teammate names
  • The link to the GitHub repository for your project.
  • Baseline inspection
    • Qualitative examples of the results obtained from the baseline method.
    • Quantitative results of the baseline method when performed by your team.
    • Quantitative results of the baseline method as reported in the paper.
  • Plans and timeline
    • A to-do list for each week leading up to the project report deadline.

Baseline Inspection

Execute the code for the baseline method and verify if there are any issues in reproducing the results as presented in the baseline paper. If pretrained models are available, run them and check whether the results align with those reported in the paper. If pretrained models are not included, you will need to conduct both the training and testing phases.

You must be able to reproduce the results in the paper using the provided code or your own implementation. If you encounter any difficulties while running the code or replicating the paper's results, you may need to consider changing the project idea. Therefore, it is strongly recommended to thoroughly examine the code of the baseline method as early as possible.

Please provide the GitHub repository link where your team will work on and push your code. This link should not be changed until the end of the project.

Plans and timeline

Provide a weekly to-do list leading up to the project report deadline. Your progress will be evaluated based on these provided timelines. Please ensure your tasks are specific and concrete. Include at least 3-5 items in each weekly to-do list.

Grading

  • Late submission: A 5% penalty for each late day in the project evaluation.
  • A submission with any missing items will be considered as not submitted.