Topic
knowledge graph
ArXiv Paper Introduces KG-SoftMAP: Bayesian Network Learning from Sparse Data Using Knowledge Graph Priors
KG-SoftMAP is a new method for Bayesian network structure learning from sparse discrete data, using a weighted knowledge graph as a prior. On synthetic benchmarks, it achieved Directed-F1 scores up to 0.97 at higher observation rates, while data-only learners stayed near zero. On real educational datasets, it matched logistic regression within 0.03 F1_FAIL while providing an interpretable concept graph.
FundaPod Introduces Multi-Persona AI Agent Platform with Knowledge Graph Memory for Fundamental Investment Research
An arXiv paper presents FundaPod, a multi-persona agent platform for AI-assisted fundamental investment research. The platform uses independent AI agents with distinct personas to gather evidence and produce testable investment memos, supported by a knowledge-graph memory system. It introduces five design principles and four architectural mechanisms for human-AI hybrid systems.
Unifying Post-hoc Explanations of Knowledge Graph Completions
Researchers from arXiv propose a unified taxonomy for post-hoc explainability in Knowledge Graph Completion (KGC). The work characterizes explanations via multi-objective optimization and discusses evaluation protocols based on metrics like Mean Reciprocal Rank and Hits@k, aiming to improve reproducibility and impact.
Initial Exploration Problem Hinders Knowledge Graph Adoption for Enterprise Users
A new paper from McNamara et al. (arXiv, 2026) theorises the Initial Exploration Problem (IEP) in knowledge graph exploration. It identifies three interdependent barriers—scope uncertainty, ontology opacity, and query incapacity—that block lay users from starting exploration. The authors argue current interfaces lack interaction primitives for scope revelation, creating a structural gap in design.
Beyond Predefined Schemas: TRACE-KG Delivers Context-Enriched Knowledge Graphs Without Fixed Ontologies
TRACE-KG is a framework that jointly constructs context-enriched knowledge graphs and an induced schema without a predefined ontology. It captures conditional relations and preserves traceability to source evidence, offering a practical alternative to ontology-driven or schema-free pipelines.
DeepRoot Multi-Agent System Enables Therapeutic Reasoning Over Historical Medical Texts with 47.6% Accuracy
DeepRoot is a multi-agent LLM system that jointly builds and utilizes a verified knowledge graph for therapeutic reasoning over historical medical texts. Applied to the Shen Nong Ben Cao Jing, it recovers 10 of 21 held-out compound-disease treatment pairs at R@20 (47.6%), significantly outperforming a raw corpus LLM (4.8%) and random baseline (2.4%). The system also reduces hallucination to 7-10% compared to 87% for tool-using LLMs, offering a scalable method for mining historical medical knowledge.
DYNA Framework Uses Temporal Knowledge Graphs to Reduce LLM Forgetting Without Retraining
Researchers propose DYNA, a lightweight framework that connects frozen large language models (LLMs) to a temporal knowledge graph, enabling continuous learning without costly retraining. On three temporal recall tasks, DYNA reduces catastrophic forgetting by ~7% compared to fine-tuning and improves temporal ordering by ~5% over standard retrieval-augmented generation (RAG). The paper also finds that higher graph clustering coefficients correlate with better retrieval, indicating the importance of graph structure.