Generative Modeling

Diffusion, flow matching, stochastic processes, discrete generation, and related sampling methods.

  1. A unified derivation of diffusion and flow matching through conditional probability paths, marginalization identities, and simulation-free regression.
    Tutorial · Published Aug 08, 2026 · Updated Aug 09, 2026 · 33 min read
  2. How continuous-time Markov chains transport categorical probability, how their rates become learnable, and how generator matching extends across modalities.
    Tutorial · Published Aug 08, 2026 · Updated Aug 09, 2026 · 33 min read
  3. How ODEs and SDEs transport probability, why scores appear in reverse-time diffusion, and how probability-flow ODEs match SDE marginals.
    Tutorial · Published Aug 08, 2026 · Updated Aug 09, 2026 · 32 min read
  4. GFlowNets from a probabilistic-ML perspective: reward-proportional sampling, training objectives, and connections to MaxEnt RL and variational inference.
    Tutorial · Published Mar 14, 2026 · Updated Aug 09, 2026 · 36 min read
  5. How path measures connect Jarzynski's equality, free-energy estimation, annealed importance sampling, diffusion models, and GFlowNets.
    Technical note · Published Mar 14, 2026 · Updated Aug 09, 2026 · 53 min read
  6. Three routes to the Fokker-Planck equation—physical intuition, heuristic discretization, and a rigorous derivation with Itô calculus.
    Tutorial · Published Feb 04, 2026 · Updated Aug 09, 2026 · 21 min read