Machine Learning for Molecules

Begin with the scientific language of molecules, then move through graph and geometric architectures, generative models, molecular simulation, proteins, genomes, and materials. Every chapter also stands on its own.

23 of 23 chapters published

Scientific foundations

  1. Chemistry and Physics for Molecular Machine Learning
    A physics-first account of molecular representation, electronic structure, forces, statistical mechanics, and dynamics for molecular machine learning.
    ML4Mol Lecture 2
  2. How Drug Discovery Turns Biological Hypotheses into Molecules
    Why drug discovery is a sequence of linked inference problems—from target validation and molecular binding to exposure, safety, and clinical benefit.
    ML4Mol Lecture 3
  3. Materials Discovery Connects Structure, Properties, and Synthesis
    A material is more than a formula: discovery must connect periodic structure, competing phases, target properties, processing, and experimental formation.
    ML4Mol Lecture 3; GDL Lecture 21

Neural architectures

  1. Graph Neural Networks as Learnable Message Passing
    Why permutation symmetry leads to message passing, how familiar GNNs instantiate it, and why graph Transformers still need structure.
    ML4Mol Lecture 4; GDL Lecture 3
  2. What Graph Neural Networks Can and Cannot Distinguish
    Graph neural network expressivity through multiset aggregation, the Weisfeiler--Leman test, its blind spots, and the cost of stronger models.
    ML4Mol Lecture 4; GDL Lecture 5
  3. Depth, Over-Smoothing, and Over-Squashing in Graph Networks
    Why deeper graph networks face under-reaching, over-smoothing, and over-squashing—and how topology determines which remedy helps.
    ML4Mol Lecture 4; GDL Lecture 6
  4. Symmetry and Equivariance for Geometric Data
    A concrete account of group actions, invariance, equivariance, and feature types for geometric machine learning.
    ML4Mol Lecture 5; GDL Lectures 1–2
  5. Scalar and Vector Geometric Graph Networks
    How geometric graph networks move from invariant distances and angles to equivariant coordinates and vector channels—and what directionality buys.
    ML4Mol Lecture 5; GDL Lecture 7
  6. Steerable Features and Tensor Products
    How irreducible rotation types, spherical harmonics, and Clebsch–Gordan tensor products create expressive equivariant neural-network layers.
    ML4Mol Lecture 5; GDL Lectures 9–10
  7. Equivariant Transformers and Machine-Learned Potentials
    How invariant attention scores, equivariant values, and energy-based force prediction turn geometric Transformers into practical interatomic potentials.
    ML4Mol Lecture 7; GDL Lectures 11–12

Generative models

  1. ODEs, SDEs, and Probability Flow
    How ODEs and SDEs transport probability, why scores appear in reverse-time diffusion, and how probability-flow ODEs match SDE marginals.
    ML4Mol Lecture 6; GDL Lecture 13
  2. Diffusion Models and Flow Matching
    A unified derivation of diffusion and flow matching through conditional probability paths, marginalization identities, and simulation-free regression.
    ML4Mol Lecture 6; GDL Lecture 14
  3. Geometric Flow Matching on Manifolds
    Flow matching beyond Euclidean space, from tangent velocity fields and geodesic conditional paths to product manifolds for molecular geometry.
    ML4Mol Lecture 6; GDL Lectures 15–16
  4. Discrete Flow and Generator Matching
    How continuous-time Markov chains transport categorical probability, how their rates become learnable, and how generator matching extends across modalities.
    ML4Mol Lecture 6; GDL Lectures 16–17

Molecular learning

  1. Molecular Data and Property Prediction Across 1D, 2D, and 3D
    How molecular representations, conformers, data splits, pretraining, and uncertainty determine what a property-prediction benchmark actually measures.
    ML4Mol Lecture 7
  2. Molecular Simulation with Machine-Learned Force Fields
    How learned energy surfaces become molecular dynamics, why rollout stability differs from static accuracy, and how to validate observables.
    ML4Mol Lecture 8
  3. Generating Molecular Graphs and Chemical Reactions
    Molecular graph generation and reaction modeling viewed as constrained structured prediction, from representation and symmetry to synthesis-aware evaluation.
    ML4Mol Lecture 9
  4. Three-Dimensional Molecular Generation and Optimization
    How generative models respect molecular geometry, how guidance turns sampling into design, and why oracle scores must survive experiment.
    ML4Mol Lecture 10
  5. Machine Learning Meets Quantum Chemistry
    Where machine learning enters electronic-structure theory, from neural wavefunctions and learned functionals to Hamiltonians and energy surfaces.
    ML4Mol Lecture 11

Biological systems

  1. Protein Structure Prediction from AlphaFold 1 to 3
    How coevolutionary constraints, pairwise geometric reasoning, residue frames, and all-atom diffusion shaped AlphaFold—and where structure prediction stops.
    ML4Mol Lecture 12; GDL Lecture 18
  2. Protein Representation Learning Across Sequence and Structure
    How sequence, alignments, residue graphs, backbone frames, surfaces, and multimodal objectives shape what protein embeddings can support.
    ML4Mol Lecture 13; GDL Lecture 18
  3. Generative Models for Protein Design
    Protein design as sequence–structure–function inference, from inverse folding and backbone diffusion to computational filters and experimental evidence.
    ML4Mol Lecture 13; GDL Lecture 19
  4. Genomic Foundation Models and Virtual Cells
    From genomic sequence models and noisy single-cell measurements to perturbation prediction and the stronger requirements of a virtual cell.
    ML4Mol Lecture 14; GDL Lecture 20