Aigarth

What is Aigarth

Aigarth is the AI running on top of Qubic and is aiming to make the third option come true. The name consists of two words, “AI” meaning “artificial intelligence” and “garth” meaning “yard” or “garden”.

Aigarth aiming to democratize AI development and enable collective intelligence. By harnessing the computational power of miners, Aigarth opens up new possibilities for collaboration and innovation in the field of AI.

Key aspects of Aigarth include:

• Decentralized AI Training: Utilizes computational power from Qubic miners to create and train billions of artificial neural networks (ANNs).
• Self-Improving AI: Implements a "Teacher" ANN that modifies and improves other ANNs, with the potential for recursive self-improvement.
• Public Accessibility: Once integrated with Qubic's Smart Contract feature, Aigarth will be publicly available for use and further development, promoting a more accessible and equitable approach to AI innovation.
• Real-World Integration: Utilizes Qubic's Oracle Machines for observing the outer world and Outsourced Computations for conducting experiments and interacting with the environment.
• Open Source: Once Aigarth is live, it will be open source for everyone.

Aigarth's integration into the Qubic ecosystem represents a significant step towards democratizing AI development and fostering collective intelligence. By leveraging the power of decentralized computing and open collaboration, Aigarth has the potential to drive groundbreaking advancements in AI and its applications across various industries.

How does Aigarth works?

It is an AI-software that get it´s information from nearly hundred thousand (number is rapidly increasing) of Qubic-miners which are creating billions artificial neural networks (ANNs) with their computational power that is supposed to compress and decompress random data. This simple task, which the ANNs are trying to accomplish, is just food for another ANN created and run by Aigarth. The latter ANN (or rather its many variations), called Teacher, modifies the former ANNs so they solve their task of compression-decompression more efficiently.

To give a simple example: One such problem could be an AI model that distinguishes cats from dogs. For this, the problem poser would provide a fitness function to rank different solutions. Participants would run a special evolutionary algorithm (a key component of Aigarth), observing in real-time how close their solution is to the perfect one.

The system's transparency allows anyone to pick up the progress of any other participant, either continuing their work or diverging towards potentially more fruitful directions. This process closely mirrors the traditional proof-of-work mining, with the notable distinction that Aigarth's "mining" has practical utility.

Under the Hood: The Helix Logic Gates

The fundamental block of Aigarth's special algorithm is the Helix logic gate. This functionally complete and reversible gate offers a simple logic: it takes three input values, A, B, and C, and outputs them in the same order after rotating by A+B+C positions. This simple mechanism enhances convergence towards an efficient solution significantly compared to randomly chosen logic gates.

Truth table and logic gate of the Helix gate