Hands-on Homeworks

We provide four hands-on problem sets plus one demo homework to help students apply the concepts learned from lectures to real chemistry problems.

Each homework contains a problem set in the pdf format, and one or more refrence Jupyter notebooks highly related to the topics covered in this homework. Students are required to turn in their answers together with source code. They encouraged to solve the problems on their own in various ways, but can refer to the notebooks for hints or possible solutions.

Homework 0 is a warm-up practice for students to get familiar with this homework form, which will not be graded:

• HW0: Get familiar with the Python environment and Jupyter notebook

  • HW0 Problem Set, HW0 Reference Notebook

  • Expected submission: Sample answers, Sample code

The four problem sets are as follows:

1. HW1: Extracting molecular descriptors

   • HW1 Problem Set

   • HW1 Reference Notebook

2. HW2: Clustering and dimensionality reduction

   • HW2 Problem Set

   • HW2 Reference Notebook

3. HW3: Supervised machine learning models

   • HW3 Problem Set

   • Reference notebooks from chapter 5 and chapter 6

     • Multi-layer Perceptron on Delaney Dataset

     • CNN on Peptide Dual-agonist Dataset

     • RNN on Delaney Dataset

4. HW4: Generative models and Protein Language Models

   • HW4 Problem Set

   • Reference notebooks from chapter 7 and chapter 8

     • VAE on QM9

     • ESM2 on Peptide Dual-agonist Dataset