Simulation of Systems of interacting mean-field Particles with High Efficiency
-
Updated
Nov 27, 2023 - Python
Simulation of Systems of interacting mean-field Particles with High Efficiency
Mean field methods in the context of continuum mechanics with special focus on orientation averaging homogenization
Solve strongly correlated condensed matter systems. Fast and simple.
Assessing consciousness by perturbing a dynamic mean-field whole-brain model fitted to empirical neuroimaging data
A python implementation of ELE method in paper "Abrupt phase transition of epidemic spreading in simplicial complexes", PRR 2020
probabilistic matrix factorization for recommender system based on gibbs sampling and variational mean field - final projects of probablistic graphical models course
High-performance optimization algorithms in Rust with Python bindings.
Research-oriented mini-projects on latent-variable modeling, approximate inference, and belief propagation in graphical models.
This repository is the official implementation of "Motif-based mean-field approximation of interacting particles on clustered networks", introducing the motif-based mean field (MMF) equations as a mean field methodology for jump dynamics on complex networks with arbitrary network motif configurations.
Dashboard documentation for the simulating the evolution of object proportions under a mean field approach.
Paper, proofs, and supporting material for global well-posedness of fully coupled McKean-Vlasov FBSDEs with jumps under full-tuple law dependence.
Exact finite-size modal coexistence and critical-endpoint scaling in the cubic Curie-Weiss model, with Lean 4 proofs and reproducible numerics.
Mean-field variational inference on MNIST, from scratch in NumPy: each digit class is a mixture of diagonal Gaussians, every coordinate update derived by hand, monotonicity of the ELBO asserted on every sweep. A companion write-up derives the method in full.
Binary pairwise MRF image segmentation over SLIC superpixels with graph-cut MAP inference and mean-field uncertainty analysis.
Certificate-first mean-field games and network equilibria: two zero-dependency interactive laboratories and a validated Python kernel. Every claim ships with a certificate.
To associate your repository with the mean-field topic, visit your repo's landing page and select "manage topics."