This repository contains my notes for the University of Bologna's Autonomous and Adaptive Systems course, held by prof. Mirco Musolesi.

An optimized LMS algorithm

AgentForge is a powerful and flexible signal-driven workflow framework designed for building intelligent, dynamic, and adaptive systems.

Classical adaptive linear filters in Julia

A professional exploration of reflective analytics, dashboards that evolve through user interaction. This project reimagines BI systems as adaptive, cognitive mirrors that learn analyst behavior, anticipate reasoning patterns, and transform data visualization into a collaborative, intelligent process.

CHI 2023 Late Breaking Work: "Exploring Physiological Correlates of Visual Complexity Adaptation: Insights from EDA, ECG, and EEG Data for Adaptation Evaluation in VR Adaptive Systems"

This project is an interactive visualization that explores the fascinating world of emergent behaviors and self-organizing systems. Inspired by the remarkable abilities of Physarum polycephalum (slime mold), this simulation demonstrates how complex patterns can arise from simple, decentralized decision-making processes.

A framework for resonance-based decision-making in artificial agents, combining celestial influences, memory feedback, and symbolic emergence.

A project for Complex Adaptive Systems' computational models

A philosophical manifesto and a series of white papers exploring a symbiotic, human-centric future for Artificial General Intelligence (AGI).

Dynamic Neural Network Refinement (DNNR) is an advanced framework that allows neural networks to adapt in real time. Unlike static systems, DNNR refines network parameters on-the-fly to optimize performance. Its modularity ensures easy customization for versatile applications.

A framework for shaping identity-aware cognition in language models using behavioral prompt layering, recursive interpretive constraints, and modular cognitive modes.

A self-healing distributed consensus framework built in Go, designed to handle network partitions, Byzantine faults, and dynamic topology changes with minimal downtime. It leverages hybrid consensus (Raft + PBFT-inspired).

🕹️🤖 Presentation for the 1st Meeting on Multidisciplinary Perspectives on Climate Change of the Resiclima Network

End-to-end toolkit for smart hash maps, CLI, Mission Control (PyQt6) dashboard, Textual TUI, workload analytics, snapshots, and automated benchmarks.

Cognition made programmable.

Python hash map that migrates between two-level chaining and Robin Hood probing based on workload. Live metrics dashboard, reproducible workloads, CSV/JSON benchmarks, safe snapshots and repair.

Samstraumr is a R&D framework in development to implement systems theory tube-based design concepts in software architecture for experimenting with adaptive systems.

Teaches LLMs to skip unproductive questions during decision-making, steering discussions toward clear specs or plans with smart tracking and focused cycles.

Sovereign AI for smart, self-learning buildings • Edge-native • 25–35 % energy savings