Merge branch 'cascadiaquakes:main' into main

Author: nts345045

README.md

@@ -1,6 +1,60 @@
 ![Deployement of book](../../actions/workflows/deploy-book.yml/badge.svg) [![Binder](https://mybinder.org/badge_logo.svg)](https://mybinder.org/v2/gh/cascadiaquakes/2025_ML_TSC/)  
-[Link to Jupyter-book template book](https://cascadiaquakes.github.io/2025_ML_TSC/)
+[Link to Jupyter-book](https://cascadiaquakes.github.io/2025_ML_TSC/)
 
-# Jupyterbook for the 2025 Machine Learning Technical Short Course
+# 2025 CRESCENT Machine Learning Technical Short Course
 
-This repository contains the notebooks for the 2025 Machine Learning Technical Short Course.
+## Program Overview
+This three-day short course provides a hands-on introduction to machine learning techniques for seismic event analysis. Participants will learn to develop AI-aided earthquake catalogs through three key steps: event detection, association, and location with quality control. The course covers neural network architecture selection, model training, performance metrics, and application to continuous seismic data. The workshop will include a mix of presentations and hands-on tutorials.  The final day will include a participant hack-a-thon in which students attempt to develop a machine learning based quality control workflow to apply to future generations of machine learning earthquake catalogs.
+
+## Learning Goals and Objectives
+
+By the end of this short course, participants will be able to: 
+
+- Explain the role of machine learning in earthquake detection, association, and location. 
+- Select appropriate neural network architectures for earthquake detection and phase picking. 
+- Train models using labeled seismic datasets and evaluate their performance. 
+- Implement trained models to detect and associate seismic events in real-world data. 
+- Optimize model parameters for accuracy and efficiency in earthquake cataloging. 
+- Integrate machine learning outputs into earthquake location algorithms. 
+- Assess model predictions and refine event catalogs through quality control methods. 
+- Design end-to-end machine learning workflows tailored to specific seismic networks or research needs. 
+- Collaborate on participant-led exercises to improve catalog quality and reliability. 
+
+##  Agenda
+
+| Time             | Day 1 (Mon)                           | Day 2 (Tue)                             | Day 3 (Wed)                              |
+|------------------|----------------------------------------|------------------------------------------|-------------------------------------------|
+| 9:00 – 10:30am   | Research Talk: AI-ready Data Set for the Pacific Northwest (Yiyu Ni, UW)             | Research Talk: Ian McBreatry gives an intro to Association             | Research Talk: Felix Waldhauser shows how to build precise earthquake catalogs             |
+| 10:30 – 11:00am  | Coffee Break                          | Coffee Break                            | Coffee Break                               |
+| 11:00 – 12:30pm  | Hack: Training a Phase Picker        | Lecture: Training a Graph Network          | Hack: Event Relocations                     |
+| 12:30 – 1:30pm   | Lunch                                 | Lunch                                   | Lunch                                      |
+| 1:30 – 2:30pm    | Lecture: Evaluating Model Performance                   | Research Talk: Multi-Geohazard Event Discrimination (Akash Kharita, UW), with a tutorial                | Science talks / Karaoke / other     |
+| 2:30 – 3:00pm    | Research Talk: Amanda talk on CNN & LFE detection                      | Research Talk: Akash Kharita presents event discrimination in the PNW                  |                                           |
+| 3:00 – 5:00pm    | Hackathon: Detect and Pick on continuous Data     | Hackathon: Establishing Quality Control Metrics?      |                                           |
+
+
+## Prerequisites  
+
+1. Participants must have intermediate python skills including: 
+
+- Core Python Proficiency – Comfortable with syntax, functions, and best practices. 
+- Data Handling – Uses pandas and NumPy for data manipulation and analysis. 
+- Automation & File Handling – Reads/writes files, automates tasks, and web scrapes with requests. 
+- Debugging & Exception Handling – Uses try-except, logging, and debugging tools. 
+- Data Visualization – Creates plots using Matplotlib, Seaborn, or plotly. 
+- Algorithms & Data Structures – Implements sorting and searching 
+- Version Control – Works with Git/GitHub, branches, and pull requests. 
+- Python Packages & Environments – Creates/imports modules, manages dependencies with venv/conda.
+
+2. Must have a laptop computer capable of accessing the internet. 
+
+
+## Instructors
+
+[Marine Denolle](https://denolle-lab.github.io/) (University of Washington)<br>
+[Amanda Thomas](https://amtseismo.github.io/) (University of California, Davis)<br>
+[Ian McBrearty](https://www.researchgate.net/profile/Ian-Mcbrearty) (Stanford University)<br>
+[Loïc Bachelot](https://loicbachelot.github.io/) (University of Oregon)<br>
+[Yiyu Ni](https://niyiyu.github.io/) (University of Washington)<br>
+[Akash Kharita](https://sites.google.com/view/akashkharita/home) (University of Washington)<br>
+[Felix Waldhauser](https://www.ldeo.columbia.edu/~felixw/) (Columbia University)<br>

_toc.yml

@@ -11,7 +11,6 @@ parts:
 - caption: Seismic Phase Detection
   chapters:
   - file: notebooks/niyiyu/curated_pnw_dataset_seisbench.ipynb
-  - file: notebooks/Loic/UNet/add_training_split.ipynb
   - file: notebooks/Loic/UNet/make_unet_annotated.ipynb
   - file: notebooks/Amanda/unet_performance_metrics.ipynb
   - file: notebooks/Amanda/apply_unet.ipynb

notebooks/Amanda/unet_performance_metrics.ipynb

@@ -5,7 +5,7 @@
    "id": "7a56d467",
    "metadata": {},
    "source": [
-    "# 🧠 GNN for Node-Level Regression: Predicting Sine Waves\n",
+    "# 🌐 GNN for Node-Level Regression: Predicting Sine Waves\n",
     "\n",
     "**Author:** Loïc Bachelot  \n",
     "**Goal:** This notebook demonstrates how to use a GNN to predict 1D sine wave signals at each node of a synthetic graph.\n",
@@ -491,7 +491,7 @@
     "- **`signal`**: Node signals (sine waves delayed by distance to a hidden origin) `[num_nodes, signal_length]`\n",
     "- **`origin`**: The hidden 2D origin point `[2]` used to generate node signals (not directly used for supervision here)\n",
     "- **`edge_index`**: Connectivity information between nodes (edges)\n",
-    "- **`edge_weight`**: Inverse-distance weights for edges (higher weight = closer nodes)\n",
+    "- **`edge_attr`**: distance weights for edges (both in x and y)\n",
     "\n",
     "This compact graph-based representation allows flexible training for node-level prediction tasks,  \n",
     "while preserving spatial and relational structure essential for learning.\n"
@@ -632,7 +632,7 @@
     "### 🏗️ Architecture Overview:\n",
     "\n",
     "1. **Signal Feature Extraction**  \n",
-    "   Each node’s raw signal is processed through a 1D convolution and max pooling to extract compact local features.\n",
+    "   Each node’s raw signal is processed through a simple MLP to extract compact local features.\n",
     "\n",
     "2. **Feature Fusion**  \n",
     "   The extracted signal features are concatenated with the node's spatial coordinates (x, y) to create a combined feature vector.\n",