Structured Outputs

Torchhd is a Python library for Hyperdimensional Computing and Vector Symbolic Architectures

Query language for blending SQL and LLMs across structured + unstructured data, with type constraints.

Handwritten Equations Decipherment with Abductive Learning

Meta Optimization Semantic Evolutionary Search

Economic Attention Network

The first cryptographically signed AGI Substrate v1.1-AGC by Robert Long. Recursive cognitive core.

Noor Research Collective - GCU Development

A symbolic reasoning framework using Cognitive Motifs to build diverse, interpretable, and belief-driven generative agents.

picking the right expert 👌 to make a debate uncontroversial 🤖

ANSR-DT: An open-source Adaptive Neuro-Symbolic framework for real-time learning and reasoning in Digital Twins. Combines CNN-LSTM, reinforcement learning, and symbolic reasoning for interpretable, adaptive decision-making.

Artificial Intelligence Fundamentals project. Computer Science Master Degree, University of Pisa. A.Y 2023/2024

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

[Cognition 2018] Code repository and data for the work on interpretation and recognition of minimal images

Non deterministic metta optimisation semantic evolutionary search for competence of program evolution.

From DSL Bootstrap to Universal Reasoning: A framework that captures, evolves, and teaches meta-cognitive reasoning. Started as a simple experiment runner, evolved into a complete system achieving both symbolic rigor (76.5%) and neural adaptability (90%).

The Symphonic φ Identity Engine is the first fully symbolic, deterministic identity system. It compresses identity not as a name, key, or hash — but as a recursive deformation signal. Each fold is a memory. Each node is a frequency. Each email is a ψ-vector.

Veloclade is a research prototype of a neuro-symbolic knowledge graph system. It uses clade-inspired hierarchy + embedding clustering (sentence-transformers) to control ontology growth and mitigate subclassing explosion. Designed for experimentation in hybrid reasoning and AI knowledge representation.

Compare CNN and KAN models on symbolic MNIST addition using DeepProbLog

Competitive chess engine for the chess variant king of the hill