Teaching

Teaching Statement

​I believe in active learning. And I believe that the role of a teacher is to facilitate students’ learning by keeping them motivated, helping students assimilate new information and knowledge into their existing cognitive schema, and creating an environment that allows them to take ownership of learning. My overall teaching goal is to prepare geoscience students for educational and career opportunities by equipping them with adequate geoscience knowledge (e.g., geology and geophysics) and computational skills (e.g., machine learning, statistics and computer programming). 

Teaching at University of Houston

Instructor. Developed lectures on potential field theory, potential field data acquisition, processing and interpretation methods.  Designed lab exercises on 3D modeling of gravity, magnetic and gravity gradient data, terrain correction, depth estimates and Fourier domain modeling. Enrollment: 37 (2020), 11 (2021), 12 (2022), 13 (2023).

Open resources: My students and I have developed an open textbook for this course. Anyone can access the book for free. This textbook is essentially a commentary of the lectures that I gave in the class. I will keep updating and adding materials as I teach this course every spring semester at UH.

Instructor. Developed lectures on finite volume method in 1D and 3D, as well as its applications to Maxwell’s equations; geophysical inversion theory and its application to EM inversions; Bayesian inversion and its application to EM inversions. Designed lab exercises on solving Maxwell’s equations using finite volume method in 1D and 3D, calculating the sensitivities, and implementing 1D and 3D EM inversions. Enrollment: 14 (2019).

Open resources:  Python codes developed for implementing 1D MT modeling, 3D DC modeling, 1D inversion and 3D sensitivity calculation using finite volume method: https://github.com/jiajiasun/GEOL6396-Computational-EM.git

Instructor. Lectured on the theory and methods for direct currents, time-domain electromagnetic, frequency-domain electromagnetic using both inductive and grounded sources, as well as magnetotellurics. Designed lab exercises using Jupyter Notebooks and Azure cloud computing. Evaluated and graded students’ homework, lab reports, final presentations and exams. Enrollment: 27 (2018), 7 (2021). 

Open resources: I have created PowerPoints slides for my lectures as well as 7 lab exercises in Jupyter Notebooks. These lab exercises allow students to interactively create a resistivity model and simulate the electrical and magnetic field as well as electrical charges. My students have found these interactive Jupyter Notebooks very useful to help them develop an intuitive understanding of the underlying physics. All the lecture slides and Jupyter Notebooks are available from my Github repository https://github.com/jiajiasun/UHElectromagnetics.git 

Co-Instructor. Responsible for (1) instructing students in the use of CG5 gravimeter and G-858 MagMapper for collecting gravity and magnetic data at Enchanted Rock and Longhorn Cavern State Park in Texas, (2) teaching students to process and interpret the measured gravity and magnetic data, and (3) evaluating and grading students’ daily reports and final presentations. Enrollment: 28 (2018), 13 (2019).

Instructor. Developed lectures on Python programming, optimization algorithms (stochastic gradient descent, mini-batch gradient descent), and several widely used machine learning algorithms such as logistic regression, support vector machine, decision trees, random forests, ensemble learning, clustering, dimensionality reduction and neural networks (including convolutional neural networks, U-net, GAN, normalizing flows and Transformer). Designed lab exercises based on real-world geoscience data in Jupyter Notebook for students to implement all the machine learning algorithms discussed in class using the Scikit-Learn package, TensorFlow and Keras. Enrollment: 23 (2018), 22 (2019), 42 (2021), 20 (2022), 30 (2023).

Open Resources: I have created PowerPoint slides as well as 11 lab exercises in Jupyter Notebooks. These lab exercises allow students to implement logistic regression, SVM, decision trees, random forest, ensemble learning, K-means clustering, PCA,  deep learning and CNN. These notebooks provide a virtual cloud environment that promotes active and experiential learning. Based on my students' feedback, this is the part that they liked the most about my course. All my teaching materials can be assessed through my Github repository https://github.com/jiajiasun/UHMachineLearning.git 

Teaching at Colorado School of Mines

Co-instructor.  Designed and gave lectures on solving nonlinear inverse problem using Gauss-Newton method, bound constrained inverse problems, general Lp norm inversions and joint inversions of multi-modal geophysical data.