Approximate Numerics in Applied Transition Environments: finite MDPs and history-dependent decision problems in Python
-
Updated
Sep 14, 2026 - Python
Approximate Numerics in Applied Transition Environments: finite MDPs and history-dependent decision problems in Python
Code for the execution experiment in master thesis "Signature-Q-Learning: A Non-Markovian Reinforcement Learning Approach""
This repository contains the codes to simulate trajectories of particles in a potential well with stochastic resetting. The dynamics are governed by generalized Langevin equation under fractional Gaussian noise.
Supplementary analysis code for the Continuity Theory Zenodo paper on seismic memory and rupture.
DARP (ICASSP 2026): online automaton refinement and planning for non-Markovian RL
This is notebook for the fractional lindblad. it serves as a demo for my paper published in zenodo (fractional master equation). it shows how much inprovements can be made if you model qubit baths from a from kernels derived from the baths geometry.
HiPO (ICASSP 2025): passive automaton learning with RPNI + online planning for non-Markovian RL tasks
Fast TCL2/TCL4/TCL6 solvers for non-Markovian open quantum systems.
compiler that treats time as a quantum state, not a parameter. Executable framework for relativistic quantum state evolution.
Non-Markovian quantum dynamics solver with a Boltzmann memory kernel (MaxEnt-derived) - a diagnostic tool for open quantum systems beyond Lindblad.
Reference implementation of all known algorithms for solving non-Markovian Reward Decision-Theoretic Planning circa 2003/04.
Neural memory diagnostics engine based on the Markov Embedding Theorem (MET). Includes a Python library, a 17-tool Model Context Protocol (MCP) server for LLMs, and a cross-platform desktop GUI.
Preprint and data for non-linear spectral structures without complexification (MSC 47J10)
To associate your repository with the non-markovian topic, visit your repo's landing page and select "manage topics."