Code of Honor

Assignments are designed to enhance your understanding and skills, constituting a significant portion of your final assessment. They must be completed individually, as engaging in any form of academic dishonesty violates the principles of the Code of Honor. If you encounter any challenges while solving the assignments, please contact the instructional team for guidance.

Use of AI for Assignments

AI may be used in an extremely restricted fashion when completing your assignments. Permitted uses include clarifying concepts, debugging errors, and performing sanity checks. However, copying AI-generated responses directly into your submission or relying on AI-generated code or solutions is strictly prohibited. The extent to which AI was used in preparing your submission must be clearly and honestly reported in the AI Usage Attestation field included in the submission page. The instructional team reserves the right to inspect submissions for potential AI usage and to apply appropriate grade deductions or other penalties if there is sufficient evidence of prohibited AI use and/or dishonesty in the AI Usage Attestation.

As mentioned in the main page, the course will consists of 5 assignment sets. Each assignment includes Written Questions that are to be solved (or briefly answered), and Programming Tasks which are to be implemented in Python.

Assignment Releases

Assignment 1: Fundamentals of Deep Learning

  • Assignment 1 has been posted on September 11, 2026
  • Deadline is on September 25, 2026 at 11:30 PM EST
  • No late submission is accepted

Assignment 2: MLPs and Optimizers

  • Assignment 2 has been posted on September 25, 2026
  • Deadline is on October 09, 2026 at 11:30 PM EST
  • No submission is accepted after the deadline

Assignment 3: CNNs and ResNets

Assignment 4: Sequence Models and Transformers

Assignment 5: Autoencoding and VAEs

How to Submit

We use Crowdmark. When submitting, please make sure to follow the instructions below. Note that these instructions are very important for correction and investigation of your submissions. Deviation from these instructions could lead to up to 10% deduction (per item).

On Crowdmark, please submit the answer of each question separately, following the steps below. Also, make sure to complete AI Usage Attestation field. Please note that failure to meet the formatting can lead to mark deduction.

  • For Written Questions, you can submit handwritten or typed answers as a .pdf , .jpg or .png file.

  • For Programming Questions, please submit a Notebook (.ipynb) file including your code and all outputs. Your Notebook should contain all figures, diagrams and codes that are requested in the question. You should make sure that the the code runs without error.

  • Please set the file name Lastname_Firstname_Assgn#_Question#.ipynb.

  • Please make sure to use the name that is displayed on the Quercus account.

  • Please make sure to complete AI Usage Attestation field, where you can explain to what extent you have used AI in completing your assignments.

The deadline is specified on the Assignment file. The deadlines are very strict and cannot be extended for the sake of fairness.