DBpedia
PulseAugur coverage of DBpedia — every cluster mentioning DBpedia across labs, papers, and developer communities, ranked by signal.
-
AI agent rewrites its own code to improve knowledge-graph question answering
A new paper published on arXiv details an AI agent capable of rewriting its own code to improve its performance on knowledge-graph question-answering tasks. This agent achieved 22% accuracy on the DBpedia benchmark, whi…
-
New benchmark ENTLORE tests latent organizational reasoning in enterprise QA · 6 sources tracked
Researchers have introduced ENTLORE, a new benchmark designed to evaluate latent organizational reasoning in enterprise question answering systems. This framework reconstructs enterprise structures from documents and or…
-
KDAI2026 lecture covers SPARQL, DBpedia vs Wikidata, and OWL
This item details lecture number 11 of KDAI2026, focusing on SPARQL interrogation tactics. The lecture covered FILTER, REGEX, OPTIONAL, UNION, negation, BIND, and GROUP BY aggregates. It also included a comparison of DB…
-
TurboVec introduces cost-efficient, private vector retrieval for enterprise RAG
Researchers have developed TurboVec, an open-source vector index designed for cost-efficient and private retrieval in enterprise Retrieval-Augmented Generation (RAG) systems. TurboVec utilizes TurboQuant, a novel codebo…
-
DBpedia Enrichment Boosts B2B Lead Recommendation Accuracy
Researchers have developed a method to enhance company representations for B2B lead recommendation systems by integrating semantic knowledge from DBpedia. This approach enriches existing company embeddings, which are ty…
-
New frameworks enhance AI-driven fact-checking against misinformation · 3 sources tracked
Researchers have developed two distinct frameworks for enhancing automated fact-checking capabilities. One, called Tree of Evidence (ToE), uses a hierarchical approach with a reinforcement learning agent to decompose, r…
-
New paper outlines principles for concept representation in sentence encoders
Researchers have published a paper detailing four principles that govern how sentence encoders represent concepts. The study, which trained encoders on millions of word pairs, found that fine-tuning primarily recalibrat…
-
New RAG Method Offers Anytime Validity for LLM Swarms
Researchers have developed a sequential extension to Federated Conformal RAG (FC-RAG) called Anytime-FC-RAG, which provides distribution-free coverage for language models at any stopping time. This new method maintains …
-
AI system learns construction materials from DBpedia without manual labeling
A project has developed an AI system capable of understanding and inferring information without explicit manual labeling. By connecting to DBpedia, the system learned to identify materials like marble, brick, and steel …
-
Developer converts monument data to intelligent semantic web
A developer has transformed textual descriptions of monuments into a semantic web, enriching them with external knowledge. This process involved applying semantic entity linking to connect monument details like dates, l…
-
COREKG method creates personalized knowledge graph summaries
Researchers have developed COREKG, a novel method for creating personalized summaries of large knowledge graphs. This approach uses coreset theory and sensitivity-based importance sampling to select a relevant subset of…
-
AI research defines low-resource languages for better digital representation
Two related papers propose methodologies to address the digital divide caused by low-resource languages in the Semantic Web. The research focuses on analyzing language distribution within Linked Open Data Knowledge Grap…
-
Researchers create Naamah, a large synthetic Sanskrit NER dataset using LLMs
Researchers have developed Naamah, a synthetic dataset of over 100,000 Sanskrit sentences designed to improve Named Entity Recognition (NER) for classical Sanskrit literature. The dataset was generated by combining enti…
-
OpenAI launches new embedding models with price cuts and performance boosts
OpenAI has released new embedding models, text-embedding-3-small and text-embedding-3-large, offering significant improvements in performance and efficiency over previous models like text-embedding-ada-002. These new mo…