Public Presentations of the Pelz Lab

Central link index for lecture decks and units.

Data Science for Electron Microscopy

• Unit 01: Intro — Slides · Source
• Unit 02: Regression — Slides · Source
• Unit 03: CNNs — Slides · Source
• Unit 04: Self-Supervised Learning — Slides · Source
• Unit 05: GANs — Slides · Source
• Unit 06: Gaussian Processes — Slides · Source
• Unit 07: Gaussian Processes II — Slides · Source
• Unit 08: Imaging Inverse Problems I — Slides · Source
• Unit 09: Imaging Inverse Problems II — Slides · Source

Mathematical Foundations of AI & ML

• Unit 01: Learning vs Data Analysis; Models, Loss Functions — Slides · Source
• Unit 02: Linear Algebra Refresher; Covariance, PCA/SVD — Slides · Source
• Unit 03: Regression as Loss Minimization — Slides · Source
• Unit 04: Neural Networks — From Neurons to CNNs — Slides · Source · Backprop self-study
• Unit 05: Clustering & Autoencoders — Slides · Source
• Unit 06: Loss Landscapes & Optimization Behavior — Slides · Source
• Unit 07: Probabilistic View of Learning; Noise; Conformal Prediction — Slides · Source
• Unit 08: Tree Ensembles for Tabular Learning — Slides · Source
• Unit 09: Latent Spaces & Advanced Representation Learning — Slides · Source
• Unit 10: Attention & Transformers — Slides · Source
• Unit 11: Generative Models — VAE & Diffusion — Slides · Source
• Unit 12: Uncertainty in Predictions — Slides · Source
• Unit 13: Physics-Informed & Constrained Learning — Slides · Source
• Unit 14: Explainability, Limits, and Scientific Trust — Slides · Source

Materials Genomics

• Unit 01: What is Materials Genomics? — Slides · Source
• Unit 02: QM Postulates, Solvable Systems, Multi-Electron Atoms — Slides · Source
• Unit 03: Quantum Chemistry Methods (HF, MP, CC, DFT) — Slides · Source
• Unit 04: Thermodynamics, Statistical Mechanics & Classical Atomistic Simulation — Slides · Source
• Unit 05: Monte Carlo Sampling & Continuum Mechanics — Slides · Source
• Unit 06: Local Atomic Environments & Universal MLIPs — Slides · Source
• Unit 07: Graph-Based Crystal Representations — Slides · Source
• Unit 08: Regression and Generalization in Materials Data — Slides · Source
• Unit 09: Neural Networks for Materials Properties — Slides · Source
• Unit 10: Representation Learning and Feature Discovery — Slides · Source
• Unit 12: Generative Models & Inverse Design — Slides · Source
• Unit 13: Uncertainty-Aware Discovery & Gaussian Processes — Slides · Source
• Unit 14: Physical Constraints, Trust, and Integration Outlook — Slides · Source

Machine Learning for Characterization and Processing

• Unit 01: What makes materials data special? — Slides · Source
• Unit 02: Physics of data formation — Slides · Source
• Unit 03: Data quality, labels, and leakage — Slides · Source
• Unit 04: From classical microstructure metrics to learned representations — Slides · Source
• Unit 05: Unsupervised methods for materials — clustering & autoencoders — Slides · Source
• Unit 06: Data scarcity & transfer learning — Slides · Source
• Unit 07: Time-series and process monitoring (W7 self-study lecture) — Slides · Source
• Unit 08: Inverse problems and process maps — Slides · Source
• Unit 09: ML for characterization signals — Slides · Source
• Unit 10: Transformers for materials characterization (ViT, Flash Attention, Mamba) — Slides · Source
• Unit 11: Uncertainty-aware regression & Gaussian Processes — Slides · Source
• Unit 12: Physics-informed and constrained ML — Slides · Source
• Unit 13: Integration, limits, and reflection — Slides · Source

Conference Talks

• 2025 MC — Slides · Source
• 2025 M&M — Slides · Source

Maintenance

The unit tables above and in index.qmd, plus each deck's "Continue" (prev/next) slide, are generated from units.yml. Edit that file, then run:

python3 scripts/build_link_tables.py

The script edits files only between <!-- BEGIN ... --> / <!-- END ... --> markers, so it is safe to run repeatedly. Use --check in CI to detect drift.

Notes

• Slides links point to the published site path on pelzlab.science.
• Source links point to editable Quarto source files in this repository.
• Print / PDF export: append ?print-pdf to any slide URL (e.g. …/01_intro.html?print-pdf) and Ctrl/Cmd+P → Save as PDF for an offline-ready archival copy. This is a built-in RevealJS feature, no extra render needed.
• Materials Genomics Units 2–4 were realigned 2026-05-09 to introduce QM, quantum chemistry, and thermodynamics content (original databases / descriptors / simulation-methods topics were redistributed across MG U6, U12, and dropped where redundant with MFML / ML-PC). See materials_genomics/REALIGNMENT_2026-05-09.md for the active map and internal/mg_realignment_old_to_new.md for historical context.
• ML-PC units use the unitNN_topic folder convention; units 9–14 keep the deck file as <NN>_<topic>.qmd rather than 01_intro.qmd.