43 results for “topic:free-energy-principle”
Research of a deep predictive coding model for unsupervised representation learning from video and audio data.
Manuscript source: Self-orthogonalizing attractor neural networks emerging from the free energy principle
Various experiments related to AGI
Simulation code and detailed exposition of the free energy princple (FEP).
A unified, computationally feasible framework for consciousness.
Generalized Predictive Coding in Torch
A lightweight SDK to calculate Predictive Integrity ( PI/∫PI )
Let's activate Net://Anchor
智能源于耗散结构优化自身不确定性的自指耗散 "Intelligence arises from dissipative structures optimizing their own uncertainty through self-referential dissipation."
Experiments and derivations for ACC2022 paper on variational message passing for online NARMAX identification.
Lean 4 library characterizing the algebraic structure of metastability and consolidation in stochastic systems
self-referential dissipative systems emerging from thermodynamic and causal constraints. | 智能源于自指耗散:从物理约束到智能涌现的自下而上生成路径
An active inference agent based on expected free energy minimization with a nonlinear autoregressive exogenous model.
Code, figures, animations for a NARX-EFE based agent.
Core mathematical functions for VIVA - sentient digital life. PAD emotions, Cusp catastrophe, Free Energy Principle, attractor dynamics.
🧮 Explore core mathematical functions for VIVA, enhancing research in emotional dynamics, consciousness, and information theory.
Biologically-inspired cognitive memory architecture for AI agents. Affect • Binding • Free Energy • Hyperbolic semantics • Metacognition.
Transforming data into intelligence, algorithms into consciousness This README is a living document. Like biological organisms, it evolves, adapts, and grows with each iteration.
This repository focused on enhancing the Active Inference Controller (AIC) for robotic manipulation tasks on an Interbotix PincherX 150 5-DOF Robot. A modified control framework called the Reactive Active Inference Controller (ReAIC) outperfors the AIC in adaptation experiments.
The Entropic Persistence Framework (EPF) is a six-layer architecture for building AI systems that persist without human caretakers. It ties entropy, energy flows, reliability, learning, embodiment, and governance into a measurable, engineering-ready model.
Bionic Self-Healing Dynamic Epigenetic Reprogramming Architecture. Intelligence requires a Dual State: The Fragile Present (Weights) and the Protected Past (DNA).
In a zero-logic ecosystem, 512 autonomous agents are tasked with navigating a environment governed strictly by thermodynamic constraints rather than predefined heuristic rules. To endure a state of chronic computational energy scarcity, these agents must evolve persistent "habits"—automated, low-entropy behavioral patterns that minimize energy loss
Online system identification in Silverbox by minimising free energy
Dark Zero Point Genesis: PPO Latent World Models Under Thermodynamic Scarcity 256 agents. 128D Latent Manifolds. Zero supervision. Agents utilize PPO-clipped surrogate objectives. Survival = Predictive Error Coding (PEC) × Energy Efficiency across a 50/15 Seasonal Cycle.
Code and experiments for IWAI 2020 submission on online system identification
Swarm dynamics based on Free Energy Principle. Dissertation project for MSc Intelligent and Adaptive Systems at the University of Sussex.
Mathematical synthesis framework unifying consciousness theories through phase transitions at brain criticality. Proposes consciousness emerges when C(t) = Φ(t) × R(t) × D(t) exceeds 8.3±0.6 bits. Integrates IIT, GNWT, FEP, and 40+ frameworks as special cases. Validated via Kuramoto simulations (qualitative agreement).
V-DV4 Active Inference: 128D Latent RSSM Manifolds. Agents minimize Variational Free Energy via 5 to 10-steps hallucinatory rollouts. Utilizing MHA Transformer Actors, behavior is optimized via BPTT through imagined time. Latent reconstruction and behavioral polymorphism under extreme thermodynamic scarcity. Zero-supervision sandbox.
Symthaea: Consciousness-first cognitive architecture — HDC (16,384D) + CfC + IIT + FEP. Live portal: symthaea.luminousdynamics.io
Your AI Homeopathic Doctor — understanding symptoms, restoring balance, and prescribing gentle remedies in harmony with your vital force.