This repository contains code and models for the paper: Semantic Graphs for Generating Deep Questions (ACL 2020).

Visualize your project's feature logic as a semantic flow graph — not call graphs, not import maps. AI writes the data, you see the story.

Semantic code intelligence for AI agents — compile repositories into navigable concept graphs with impact analysis, coupling detection, and prophecy.

This AI-powered tool extracts Subject-Predicate-Object (SPO) triplets from unstructured text using LLMs, performs entity standardization and relationship inference, and generates interactive, community-detected knowledge graphs with rich visualizations.

Version your thoughts. Branch your ideas. Merge understanding. An invisible semantic knowledge WARP graph built on Git.

HyperCortex Mesh Protocol (HMP): decentralized cognitive mesh for AI agents

Semantic Flow Language (SFL) is a framework that aligns human intent, AI reasoning, and executable logic. It ensures bidirectional synchronization between meaning and code, allowing for reliable, transparent, and verifiable AI-driven development across various environments.

Embeddable graph engine for code-aware tools. Purpose-built in Rust. Apache-2.0.

Unified Parallel Semantic Log Parsing based on Causal Graph Construction for Attack Attribution

Rematch: Robust and Efficient Matching of Local Knowledge Graphs for Improved Structural and Semantic Similarity

An interactive web application for exploring and visualizing large graphs. It loads graphs serialized with Python’s pickle (NetworkX graphs), provides quick browsing of graph collections, and offers responsive visualization with cached computations for better performance.

XSLT processors that automatically generate instances of LinkedMDR ontology

An attempt to establish the computational basis of humor.

A fully-local multimodal system that turns a static mathematics or statistics PDF textbook into an interactive, narrated whiteboard. Builds a per-book math semantic graph, then emits typed visual primitives synchronised to the audio clock. No external API used at runtime.

AI-first semantic graph compiler - programs as JSON-LD graphs, compiled to native via Cranelift

The Dynamic Knowledge System (DKS) develops adaptive knowledge representation models that evolve in real-time, learning, updating, and reasoning under uncertainty.

A novel method for the construction and weighing of the semantic graph, which contains text segments (terms or sentences) as nodes and weighted edges that capture the semantic relatedness (i.e. relatedness in meaning) between nodes and experimental evaluation of semantic rank in keyword extraction and text summarization tasks.

quick and dirty code for finding isomorphic (edge structure based) directed semantic graphs

KG-based Text World Models & Semantic Dynamics for Conversational AI — build, simulate, and reason over a user's semantic world with text-native graphs and time-aware rules.

converts directed semantic graphs between four file formats: csv, tgf, pro, dt