Machine Learning for the Sciences
The material in this repo is part of a seminar at Princeton University. Feel free to use it as you like.
Approach
The seminar started out with several lecture-style meetings on the fundamentals of machine learning and data science. From there, it covers code demonstrations, research projects, and paper discussions from recent literature.
The "Lecture"
- Introduction and Introductory Example
- Kernel density estimation
- (Gaussian) mixture models
- Mixture model applications
- Clustering
- Classification overview and Theory for linearly separable cases
- Neural networks 101 and PyTorch introduction
Tutorials / Research
- Flux estimation and its priors (Jim Bosch)
- A (non-traditional) introduction to TensorFlow (Dan Foreman-Mackey)
- Likelihood-free inference (Justin Alsing)
- Sparsity and Deep Learning for modern surveys (Francois Lanusse)
- Practical Hamiltonian Monte Carlo in Python for Astronomers (Dan Foreman-Mackey)
- Forward, causal modeling of galaxy photometry: Hierarchical causal models ↔ machine learning (Boris Leistedt)
- Using deep convolutional neural networks to predict galaxy metallicity from three-color images (John Wu)
- Accelerating Scientific Calculations with GPUs (Evan Schneider)
Papers / Blogs
- The Building Blocks of Interpretability by Olah et al. (2018)
- Galaxy detection and identification using deep learning and data augmentation by González et a. (2018)
- The Unreasonable Effectiveness of Recurrent Neural Networks by Andrej Karpathy (2015)
- Understanding LSTM Networks by Christopher Olah (2015)
- Robort Burns by Adrian Price-Whelan
- The fall of RNN / LSTM by Eugenio Culurciello (2018)
- Deep neural networks to enable real-time multimessenger astrophysics by George & Huerta (2018)
- Galaxy detection and identification using deep learning and data augmentation by González et al. (2018)
- Show, Attend and Tell: Neural Image Caption Generation with Visual Attention by Xu et al. (2015)
- From Dark Matter to Galaxies with Convolutional Networks by Zhang et al. (2019)
- Hamiltonian Neural Networks by Greydanus et al. (2019)
- Differentiable Convex Optimization Layers by Agrawal et al. (2019)
Conferences
- Interpretable Learning in Physical Sciences October 14 - 18, 2019
- Theory of Deep Learning: Where next? October 15 - 18, 2019
- Machine Learning and the Physical Sciences December 14, 2019