Projects
Fraktalia & Open-Source Initiatives
Featured Project: Fraktalia
Fraktalia (historically developed under the project name openEugenia) is an open-source, educational, and social platform that unites personal knowledge management, peer-to-peer learning, and human-centric Artificial Intelligence.
The project is built on two core dimensions: a personal dimension, where users build and reinforce their sovereign knowledge base, and a social dimension, where users connect around shared and complementary knowledge.
Architectural Philosophy & Data Sovereignty
Fraktalia embraces the movement toward a decentralized web (Web 3.0):
- Peer-to-Peer Connections: Users connect directly through a P2P architecture without relying on a central server that holds or monetizes their personal information.
- Full Data Sovereignty: When software manages data as intimate as an individual’s knowledge base, it must remain private and completely under the user’s control.
- Ethical, Non-Commercial AI: Fraktalia’s algorithms and assistants operate strictly without commercial data exploitation, ensuring total user autonomy.
The Personal Dimension: Knowledge Base & Eugenia
1. Multi-Format Knowledge Extraction
Users build a personal database composed of individual knowledge items extracted from diverse multi-modal sources, including scientific papers, PDFs, web pages, images containing text, and audio/video transcripts.
To remove the burden of manual formatting, Eugenia—the user’s personal AI agent—automatically detects and suggests key knowledge snippets worth preserving from submitted documents.
+-------------------+
| Multi-Modal | (PDFs, Web pages, Images, Audio/Video)
| Document Sources |
+---------+---------+
|
v
+-------------------+
| Eugenia Agent | (Automated natural language extraction)
+---------+---------+
|
v
+-------------------+
| Personal | (Database of structured knowledge snippets)
+-------------------+
2. Active Knowledge Reinforcement & Spaced Learning
Preserving knowledge requires active recall. Fraktalia integrates an adaptive learning algorithm:
- Mastery Score (\(s\)): Every piece of knowledge is assigned a score from \(0\) (low mastery) to \(1\) (high mastery).
- Interactive Quizzing: Eugenia conducts conversational review sessions on topics with lower scores, recalculating the mastery level after each answer based on semantic comparison.
- Temporal Aging Factor (\(f\)): To account for natural memory decay over time, scores decay periodically; however, as a concept is repeatedly mastered, its stability increases and decay slows down.
Getting Involved & Code Repository
Fraktalia is an open-source project welcoming developers, researchers, and educational scientists.
- Source Code: Hosted on our official openEugenia GitHub repository.
- Contributions: We invite contributions in P2P protocols, natural language processing, privacy-preserving algorithms, and UI/UX design.