Python is a great programming language that is relatively easy to learn and also very easy to read.
The data we are using for this lesson are from the Portal Project Teaching Database
More details about the files we’ll use and where to download them are available on the Setup page
Prerequisites
Learners need to understand the concepts of files and directories (including the working directory) and how to start a Python interpreter before tackling this lesson. This lesson references the Jupyter notebook although it can be taught through any Python interpreter. The commands in this lesson pertain to Python 3.
To get started with installing Python, follow the directions given in the Python section of the course Software page. In addition to installing Python on your own computer, you will also need to download the data files used in the tutorials. Details for doing this are found in the Setup page.
| Data Sources |
Understanding data sources
How to get data from online sources How to retrieve dataset with the Toolbox? |
|
| Data Preprocessing and Visualization | How do I processing nc data? | |
| Machine Learning Fundamentals |
What are the fundamental concepts in ML I can use in sklearn framewrok ?
How do I write documentation for my ML code? How do I train and test ML models for Physical Sciences Problems? |
|
| Deep Learing Fundamentals |
What are the basic timeseries I can use in pandas ?
How do I write documentation for my Python code? How do I install and manage packages? |
|
| Recurrent Neural Networks (RNNs) Models |
What are the primary advantages of using Recurrent Neural Networks (RNNs) for time series forecasting over traditional statistical methods and other machine learning algorithms?
What are the key differences between traditional RNNs and advanced RNN models such as LSTMs and GRUs? What are some common challenges faced when training LSTM models and how can they be mitigated? How do Long Short-Term Memory (LSTM) networks and Gated Recurrent Units (GRUs) enhance the capability of RNNs in learning and remembering temporal dependencies in sequential data? What recent advancements in RNN variants, such as the Temporal Fusion Transformer (TFT), have contributed to improved time series forecasting in physical sciences applications? |
|
| Machine Learning For Material Science |
Can we have a comprehensive dataset consisting of large structures from various compounds spanning the whole periodic table
Do we have an machine learning libary that able to study electron interactions and charge distribution in atomistic modeling with near DFT accuracy |