Pure MeTTa implementation of MetaMo (Lian & Goertzel 2025) on the sivaji1012/Core MeTTa engine.
Equation #9 — incremental self-model blend:
x_{t+1} = (1 - α) · x_t + α · x*
α = α_0 · (1 - g_over^Ind) + β_0 · g_over^Trans
Where x = (G, M):
G = (g₁, g₂, …)— goal intensity vectorM = (μ₁, μ₂, …, μ₆)— modulator vector (valence, arousal, approach, resolution, threshold, securing)
Lian & Goertzel (2025). MetaMo: A Robust Motivational Framework for Open-Ended AGI.
AGI 2025 Proceedings (LNAI 16057), Chapter 34, pp. 386–398.
!(import! &self (library lib_metamo))
!(bind! &ms (new-space))
!(MetaMo.seed! &ms my-goal (G 0.8 0.3) (M 0.5 0.4 0.6 0.7 0.3 0.8))
!(MetaMo.update! &ms my-goal (G 0.6 0.7) (M 0.3 0.6 0.4 0.5 0.7 0.6))
!(MetaMo.get &ms my-goal)
; → (MotiveState my-goal (G 0.78 0.34) (M 0.48 0.42 0.58 0.68 0.34 0.78))| Function | Description |
|---|---|
(MetaMo.seed! &space goal G M) |
Initialize a goal's MotiveState |
(MetaMo.get &space goal) |
Read current MotiveState |
(MetaMo.update! &space goal G* M*) |
Apply equation #9 blend |
(MetaMo.alpha g-ind g-trans) |
Compute blend rate α |
(MetaMo.goal-strength &space goal i) |
Read G[i] |
(MetaMo.modulator &space goal i) |
Read M[i] |
(MetaMo.stats &space) |
Summary stats |
# Batch (single JIT warmup)
./test.sh
# Interactive (Revise.jl — instant reruns)
julia --project=../Core
julia> using Revise, MeTTaCore
julia> include("repl_test.jl")
julia> run_metamo_tests()Use (with-space &scratch (new-space) body...) for transient evaluation
contexts — releases the space reference on exit, preventing leaks.
Only bind! spaces intended to be long-lived.