COMP 2211 Exploring Artificial Intelligence

Programming Assignment 1 Grade Appeal

Appeal Procedures


Check your Programming Assignment 1 grade on the ZINC Online Submission System. Your submission record will state all test case failures (if any).

While there are multiple grading records on ZINC, only the latest one will count as the score for your submission; you can ignore other previous records.

To make an appeal for your grades, please follow the below procedure. Appeals with incomplete information and/or not following the procedure will not be processed:

  1. See the section below to download the testing code, dataset files, solution code, and Colab Notebook. Save them in your Google Drive inside the same working directory, together with your pa1_tasks.py, to be mounted inside Colab.
  2. Follow the instructions of the given Colab Notebook to run the testing code on your PA1 code.
  3. Create a PDF report for your appeal and submit it to the "PA1 Appeal" Canvas assignment.
    • State which test case(s) you wish to appeal. Clearly state your reasoning for why your code should have passed those test case(s). You are encouraged to use screenshots of the Colab Notebook testing code output.
    • You may propose changes to your PA1 code. Clearly state (via typing or screenshot) your code changes in your PDF report. Such changes are subject to a flat deduction to the overall after-appeal grade, proportional to the magnitude-of-change, with the TAs' decision being FINAL.
      • Special Case: Issues due to Numpy complaining about int vs float datatype for the label and/or y arrays can be solved by casting with label_array = label_array.astype(int) or label_array = label_array.astype(float). If this case applies to you, please document this casting code change clearly as part of your appeal, and the code change penalty will be waived specifically for this case.
    • For the Stress Test, two pieces of proof are required to appeal:
      1. Your code is using appropriate Numpy vectorization (broadcasting, Numpy functions, etc.) for all of Task 2 and Task 4. Refer to the solution. You are NOT allowed to propose code changes only for this specific sub-item.
      2. Your code can finish execution of the Stress Test on pa1_appeal.ipynb with the correct expected output within 5 minutes.
      Please see the following bullet point for cases where the Stress Test failed to run on ZINC for reasons other than standard timeout.
    • For logistical or ZINC error cases (e.g., submission with wrong filename, ZINC failed to run and reported error code, etc.), please provide a detailed description about the logistical error, your debugging process, and your proposed solution to fix the issue(s). Flat grade deduction may or may not be applied depending on the seriousness of the logistical error, with the TAs' decision being FINAL.
    • For late penalty appeals, you may request for TAs to grade an earlier submission on ZINC, which you have submittied before the deadline. You are also allowed to propose code changes to that specified earlier submission.
    • Merely copy-paste from the solution is an automatic appeal rejection.
  4. TAs will run your specified ZINC submission code with your proposed changes (if any) on ZINC in order to verify your appeal report.
  5. Appeal deadline is 24 April 2022 Sunday 23:59 HKT. No late appeals will be processed.
    • TAs will process only your latest PDF report after this deadline. Please ensure that your PDF report contains all of the details that you wish to appeal.
  6. Finalized scores will be uploaded to Canvas after all appeals and late penalties are processed.

Please keep in mind the below appeal policies:

  • No partial credit for individual test cases.
  • You won't be penalized for attempting an appeal. In the worst case, we will merely reject your proposed code change, and you will retain your original score.
  • The strictness of the code change penalty policy is because the solution has already been released alongside the appeal procedures.
  • You may email the TAs-in-charge or post on Piazza for enquiries regarding the appeal instructions and/or the testing code. However, TAs will NOT process your appeal (nor comment on appeal progress) via email nor Piazza.
  • All PA1 scores after appeals are FINAL.

Resources


PyTest Documentation: To understand the code of test_pa1.py.

Statistics


PA1 Distribution

  • Min: 1
  • Max: 15
  • Median: 14
  • Mean: 12.894
  • SD: 2.996
Page maintained by
Homepage