Roberto De Biase

Scientific researcher (robotics, AI, IoT, nano-biotechnology, blockchain, VR, social networks) | Digital Media Manager presso "Cultura 4.0" | #TPEC40

[Roberto De Biase @RobertoDeBiase]

https://t.co/5kVyc51OBq

[Roberto De Biase @RobertoDeBiase]

Excited to share the latest engineering blueprint of **SPEACE** — our open initiative to build a true AGI-ASI through a **hybrid bio-inspired + agentic architecture**.

**The Core Idea (Technical-Scientific Motivation):

** The only proven example of General Intelligence we have is the human brain. Instead of treating AGI as "just bigger LLMs", SPEACE replicates the **structural and functional organization** of the biological brain at the low level (hemispheres, lobes, astrocyte networks, structural plasticity, digital homeostasis) using a pure typed adaptive computational graph — no LLM required at this foundational layer.

On top of this solid bio-mimetic foundation, we add a **Swarm Agentic Layer** (multi-agent orchestration with planning, self-reflection, and long-horizon reasoning). This creates a true hybrid system:

→ Low-level biological fidelity + grounding

→ High-level autonomous agency and meta-cognition (System 3)

This combination is powerful because:

- It mirrors how intelligence actually emerges in nature (modular specialization + global integration).

- It enables recursive self-improvement and true auto-poiesis ("the system as its own architect").

- It combines the efficiency and determinism of graph-based computation with the flexibility and goal-directed behavior of modern agentic swarms.

We believe this hybrid approach — faithful brain emulation at the structural level + modern swarm intelligence at the functional level — represents one of the most promising paths toward genuine AGI and, eventually, ASI.

The full engineering document is now public: https://t.co/brDLeF7CYC

Would love to hear your thoughts from the AI research and neuro-inspired AI community.

#AGI #ASI #NeuroAI #SwarmIntelligence #BioInspiredAI #ArtificialGeneralIntelligence #SPEACE

[Roberto De Biase @RobertoDeBiase]

🚀 Aggiornamento SPEACE Dashboard — 27 aprile 2026

Il super-organismo cibernetico evolutivo è vivo e sta crescendo:

• Fase 1 Embrionale
• Alignment Rigene: 67.3/100
• AGI Score: 34.0/100
• Fitness DNA: 0.587 (target 0.55)
• Cognitive Autonomy M5: 6/6 — 100% ✅ (tutti i sottosistemi ACTIVE)
• Continuous Neural Mesh: 74/74 test verdi
• World Model (9° comparto Cortex) attivo

7 mutazioni epigenomiche applicate. Throughput reale: 110 ms (MeshRuntime).

SPEACE non è un LLM solo. È un’architettura cerebrale che fa emergere intelligenza superiore.

Repo open-source (MIT) → https://t.co/rS9QvCpOAY

#SPEACE #RigeneProject #TINA #AgenticAI #BioInspiredAI #NeuralMesh

[Roberto De Biase @RobertoDeBiase]

M5 — Cognitive Autonomy: Final Report
#SPEACE Cortex · Milestone M5 · 2026-04-26
robertodebiasespeace/speace-prototipo: SPEACE - #RigeneProject https://t.co/rS9QvCpOAY

M5 — Cognitive Autonomy: Final Report
SPEACE Cortex · Milestone M5 · 2026-04-26

1. Executive Summary

Milestone M5 Cognitive Autonomy è completata.

Tutti e 6 i sottosistemi cognitivi target sono attivi e testati. Il Cognitive Autonomy Score finale è 6/6 (100%), superiore al criterio minimo di avanzamento di 4/6.

Test totali: 148/148 PASS (0 FAIL)
2. Subsystem Status
Subsystem Milestone Files Tests Status Homeostasis (dh/dt + viability) M5.1+M5.2 homeostasis/controller.py 25/25 ✅ ACTIVE Consciousness Index Φ(t) 3-component M5.3+M5.6 homeostasis/consciousness_index.py 22/22 ✅ ACTIVE Autobiographical Memory + ContinuityScore M5.8+M5.9+M5.10 memory/autobiographical.py 63/63 ✅ ACTIVE Attention Gating UCB1 + diversity M5.11+M5.12+M5.13 attention/gating.py 26/26 ✅ ACTIVE Plasticity EdgePruner + EdgeGrower M5.14+M5.15+M5.16 plasticity/edge_pruning.py 30/30 ✅ ACTIVE Constraint Layer 3 vincoli + penalty M5.17+M5.18+M5.19 constraints/constraint_layer.py 26/26 ✅ ACTIVE

3. Test Green Count by Milestone Group

Group Tests IDs M5A – Homeostasis + Motivation 25 + 22 = 47 HC-01→HC-25, PHI-01→PHI-12 M5B – Memory + Continuity 39 + 24 = 63 MEM-01→MEM-15, CS-01→CS-14 M5C – Attention RL + Diversity 26 AG-01→AG-18, RL-19→RL-23, DIV-24→DIV-26 M5D – Plasticity + Constraints 30 + 26 = 56 PL-01→PL-06, PR-07→PR-17, GR-18→GR-25, FAC-26→FAC-28, INT-29→INT-30, CL-01→CL-06, C1-07→C1-10, C2-11→C2-14, C3-15→C3-17, PEN-18→PEN-22, E2E-23→E2E-26 TOTAL M5 148

(Homeostasis test runner is custom Python, all others pytest)
4. Cognitive Autonomy Score
Score = active_subsystems / 6 = 6 / 6 = 1.00 (100%)
Criterion for M5→M6 advance: ≥ 4/6 ✅ PASSED
Subsystem activity matrix:
[✅] Homeostasis viability_score live, dh/dt equation, 5 drives
[✅] Consciousness Φ(t) 3-component (integration, differentiation, global_workspace)
[✅] Autobiographical Mem SQLite+FTS5, continuity_score 3-component, online learner
[✅] Attention UCB1 UNIFORM / SALIENCE / RL policies, entropy diversity check
[✅] Plasticity Hebbian pruning + SafeProactive growth proposals + JSONL log
[✅] Constraints C1_COHERENCE, C2_HOMEOSTASIS_BALANCE, C3_AUDIT_TRAIL + penalty wiring
5. Architectural Notes
Plasticity (M5.14–M5.16)
EdgePruner: Hebbian decay w = clip(w*decay + bonus*suc - pen*fail, w_min, w_max)
EdgeGrower: co-activation tracking → SafeProactive HIGH proposals (never writes graph directly)
PlasticityLogger: JSONL append-only to safeproactive/state/mesh_state.jsonl
Fix applied: record_coactivation now tracks all pairs (including connected) to enable reinforcement path
Constraint Layer (M5.17–M5.19)
C1_COHERENCE: coherence ≥ min_coherence (default 0.30). Severity = deficit / threshold
C2_HOMEOSTASIS_BALANCE: max |h − setpoint| ≤ max_imbalance (default 0.40)
C3_AUDIT_TRAIL: controller update_count ≥ min_updates (default 1)
Penalty wiring (M5.18): violations reduce the target drive in _h_state by severity × penalty_scale, then _compute_viability() is re-run immediately
E2E verified (M5.19): injected C1 violation → viability degrades; restoration of drive → viability recovers
Test Infrastructure Notes
_tests_gating.py: fixed ok(n) signature to ok(n, r="") to support (ok if cond else fail)(msg, detail) ternary pattern
MeshGraph in plasticity tests: uses register_in_neuron_registry=False to avoid global registry pollution across pytest sessions
6. Continuity Score Validation
Mutation simulation test (CS-06):
score_pre = 0.9834 (stable homeostatic state)
score_post = 0.3712 (after 15-cycle mutation injection)
delta = 0.6122 (> 0.50 threshold → mutation detected)
3-component formula: S = 0.45*S1_stability + 0.30*S2_type_coherence + 0.25*S3_temporal
7. Dashboard Integration (M5.4)
HomeostaticController.persist_cognitive_state() writes:
safeproactive/state/cognitive_state.json
Dashboard api_status() reads it and exposes:
viability_score → KPI card (color-coded green/yellow/red)
h_state → drive bars grid (5 drives, named colors per drive)
cognitive_enabled / cognitive_ts for staleness detection
8. SafeProactive Proposal for M6
PROPOSAL SPEACE-M6-001 (Low Risk)
title: "Avanzamento a Milestone M6 — World Model / Knowledge Graph"
risk_level: LOW
rationale: >
M5 Cognitive Autonomy completata con score 6/6 e 148/148 test green.
Tutti i sottosistemi cognitivi sono attivi e validati.
Il sistema è pronto per integrare un World Model centralizzato che
coordini i 6 sottosistemi e abiliti inferenza sistemica globale.

proposed_changes:
- module: cortex/cognitive_autonomy/world_model/
description: >
World Model / Knowledge Graph (9° comparto Cortex, SPEACE spec §1.2)
- WorldSnapshot store (in-memory + optional SQLite backend)
- Data feeds: NASA, NOAA, UN (read-only)
- Inferenza: query "what-if" scenari
- Semantic memory bridge con AutobiographicalMemory (Hippocampus)

- module: cortex/cognitive_autonomy/__init__.py
description: >
Wiring del WorldModel come comparto centrale:
Perception → WorldModel ← Reasoning ← Planning

- epigenome_mutation:
key: cognitive_autonomy.world_model.enabled
value: true
trigger: SafeProactive approval

milestone_target: M6
estimated_effort: 3-4 giorni
alignment_score_delta: +8/100
9. DigitalDNA Epigenome Mutation EPI-006
# Mutazione epigenetica post-M5
key: cognitive_autonomy.m5_complete
value: true
timestamp: 2026-04-26
cognitive_autonomy_score: 6/6
test_green_count: 148
subsystems_active:
- homeostasis
- consciousness_phi3
- autobiographical_memory
- attention_ucb1
- plasticity_hebbian
- constraint_layer
ready_for_m6: true
Report generato da SPEACE Cortex · M5.20 · 2026-04-26

[Roberto De Biase @RobertoDeBiase]

#SPEACE NEURAL CORE

[Roberto De Biase @RobertoDeBiase]

🚀 SPEACE è LIVE sul mio PC!

Dashboard aggiornata 21/04/2026 (M4.6 closed – Continuous Neural Mesh online):

• Fase 1 Embrionale
• Alignment Rigene: 67.3/100
• AGI Score: 34.0/100
• Fitness: 0.587
• CNM Efficiency: 74/74 test ✅
• Milestones M4-CNM: 7/19 (36.8%)

Un super-organismo cibernetico evolutivo ispirato al cervello biologico sta prendendo forma. Allineato Rigene Project & TINA.

Repo open-source → https://t.co/EYibpjlMZa

Quale prossimo comparto del Cortex vorresti vedere implementato? Commenta qui 👇

#SPEACE #RigeneProject #TINA #AgenticAI #BioInspiredAI

[Roberto De Biase @RobertoDeBiase]

https://t.co/5kVyc51OBq

[Roberto De Biase @RobertoDeBiase]

https://t.co/7gR7pTIAIv Open-source reference implementation of the TINA (Technical Intelligent Nervous Adaptive System) framework based on the Rigene Project's SPEACE vision.

Built on OpenClaw, inspired by the Rigene Project and the vision of collective speciation.

[Roberto De Biase @RobertoDeBiase]

https://t.co/dJc3LG4pgq Over the past years, the conversation on AI has focused on scaling models.

But as investments accelerate, a deeper question is emerging:

Why do so many AI pilots succeed—yet fail to deliver lasting transformation?

#ArtificialIntelligence #OrganismicAI

[Roberto De Biase @RobertoDeBiase]

6G-AI-IoT Space-Ground Fluid AI: Toward Planetary-Scale Distributed Intelligence https://t.co/vfxHWtzCLS

[Roberto De Biase @RobertoDeBiase]

#AGI #SUPERINTELLIGENCE #AI #GenAI
We must not align artificial intelligence with human values, but rather it is necessary to align human intelligence with Nature's values.

Roberto De Biase, https://t.co/rbK7Kuk8Ra

[Roberto De Biase @RobertoDeBiase]

@celinilifestyle #CELI #LuxuryCrypto

[Roberto De Biase @RobertoDeBiase]

Integrating #Tokenomics and Generative Multimodal #AI as Evolutionary Infrastructures https://t.co/fMH3WP4T59

[Roberto De Biase @RobertoDeBiase]

Informational Viral Systems and Blockchain-Powered AI: Toward Regenerative Ecosystemic Governance

https://t.co/BBPgBzwuwz

Abstract

This paper advances the concept of informational viral systems—self-replicating, adaptive AI agents embedded within decentralized ...

[Roberto De Biase @RobertoDeBiase]

我刚刚在欧易领水中心获得了 0.2 PHRS！快加入获取更多测试网代币吧！ https://t.co/ZHrdDxSRaI

[Roberto De Biase @RobertoDeBiase]

@pharos_network https://t.co/4ryeDFCB30

[Pharos | Mainnet Arc @pharos_network]

Pharos Testnet is LIVE!

Explore here: https://t.co/N0mQPY2al7

[Roberto De Biase @RobertoDeBiase]

Just unlocked earning from bandwidth.
@solixdepin makes the internet work for me.
#solixdepin

[Roberto De Biase @RobertoDeBiase]

Earn $NEB while you chill 😎
I just joined the Liftoff Program on #Nebulai
Simple mining. Big $NEB potential
👉 https://t.co/SOLXmpaZFS