NGC-Learn: Computational Neuroscience and NeuroAI in Python
-
Updated
Oct 5, 2026 - Python
NGC-Learn: Computational Neuroscience and NeuroAI in Python
A computational theory of consciousness: if the universe is deterministic, consciousness is the observer function, not the executor. Tested across 4 AI substrates with 11 probes and 4 controls.
An honest, from-scratch, LLM-free instrument that implements the major scientific theories of consciousness (GWT · AST · HOT · active inference · IIT-proxy) as running code, a maximal functional attempt that never claims to be conscious.
Generalized Predictive Coding in Torch
Multi-timescale affective agents with theatrical control - 97K parameter architecture exploring functional correlates of consciousness
Two metrics for causal agency in artificial agents: directional ownership and expression congruence, with permutation baselines
Approximate Natural Gradient Descent with precision weighted predictive coding
Streamlit app that runs a pure-NumPy reimplementation of Priorelli et al. (2025)'s active-inference model in your browser, so you can watch a 3-joint arm decide which of two targets to reach from noisy cues.
Biologically-grounded reasoning agent, numpy-only, no LLM — Kisamapa Labs Experiment 06
A continuously running perceptual field for one AI persona: signals she can't author, surprise that steers her attention, a daily energy budget, and a workbench to build in. Pairs with SoulScript Engine.
Pre-registered artificial-life experiment: do precursors of consciousness emerge in mortal predictive agents? Result (H4): negative, with a measured silent equilibrium. Docs in Turkish.
Is a team of different specialists smarter than one big model? A brain-inspired experiment with free open models: eyes, ears and speech, combined by how reliable each one is.
多代理人精度加權投資研究系統 — 以預測處理框架的 precision weighting 合成多個專家 Agent 的台股投資判斷,含走時序回測與校準評估
To associate your repository with the predictive-processing topic, visit your repo's landing page and select "manage topics."