Loading...
Loading...
Social network graphs, knowledge graphs, citation networks, molecular graphs for GNN, web link graphs
322 datasets
VibeSearchBench is a dataset containing 200 simulated research tasks for evaluating search systems. It includes 100 daily-life tasks and 100 professional tasks, each with a ground-truth knowledge graph. The dataset was created by VibeSearchBench and was last updated on May 6, 2026.
Comparison data evaluates HP-GNN model performance across three datasets, including results on 22 pediatric patients. The dataset was authored by Masoud Amiri and last updated in April 2026. It contains evaluation metrics in a 5.5 KB Excel file.
Python source code implements a Hypergraph Neural Network (HP-GNN) model for seizure prediction. The package includes model architecture, training scripts, and evaluation tools for data from 22 pediatric patients. It was authored by Masoud Amiri and last updated in April 2026.
5.5 KB of tabular data compares the computational demands of a Hypergraph Neural Network against baseline methods for seizure prediction. The dataset, created by Masoud Amiri and shared on figshare, was last updated in April 2026. It originates from research involving 22 pediatric patients and Kuramoto oscillator dynamics.
Cybersecurity RAG Knowledge Graph is a structured, synthetic dataset designed for retrieval-augmented generation systems and AI copilots. The preview contains a knowledge graph built from 75 articles organized into 25 topics, resulting in approximately 200 text chunks. The dataset was created by Lucasautomatekc and last updated on Hugging Face in April 2026.
5.5 KB of benchmark results for inductive knowledge graph completion, measuring HITS@10 scores. The data, published by Qingsong Li on figshare, compares multiple methods across four standard inductive split versions (v1-v4). Results were last updated on March 19, 2026.
Wind speed data averaged within 25x25 kilometer grid cells along satellite tracks, derived from the CYGNSS constellation's Delay Doppler Mapping Instrument. NOAA/NESDIS produced this version using a specific geophysical model function and track-wise debiasing algorithm. One netCDF-4 file is generated per day, containing data from up to eight spacecraft with a latency of approximately six days.
Curated papers, datasets, and benchmarks for social network simulation, from classical network models to LLM-based agentic social systems. The repository is authored by tamlhp and was last updated on 2026-05-07 11:53:09.
RiTeK is a benchmark for complex reasoning over medical Textual Knowledge Graphs (TKGs). It contains three medical graph QA subsets, including ADint, to evaluate retrieval systems and Large Language Models (LLMs). The dataset was created by ChenAI2015 and last updated on 2026-04-11.
Replication data hosted for the paper 'Are LLM-Enhanced GNNs Privacy-Safe?' by Jimmy Wen. The dataset likely contains graph-structured data used to evaluate privacy risks in hybrid LLM-GNN models. It was last updated on May 6, 2026.
9.5 KB of benchmark results for transductive knowledge graph completion models, including NBFNet and LMKE. The data was compiled by Qingsong Li, with results taken from original papers and some trained with original experimental settings. The dataset was last updated on March 19, 2026.
EvidenceNet provides structured knowledge graphs derived from approximately 500 full-text biomedical articles per disease, published between 2010 and 2025. Created by Chang Zong and released via Harvard Dataverse, this dataset transforms literature into evidence records with normalized entities and quality scores. The public release includes two disease-specific resources: EvidenceNet-HCC and EvidenceNet-CRC.
Hongtu Zhu's Data Management and Sharing Plan outlines the scientific data strategy for constructing comprehensive knowledge graphs for Alzheimer's disease research. The plan describes the data to be generated and used, with a focus on managing and sharing project data. It was last updated on April 27, 2026.
OS MasterMap® Highways Network is described as the most complete, detailed, and accurate navigable road network dataset for Great Britain. It records road dimensions and accessibility, drawing on authoritative sources to support operational decisions. The dataset includes information on planned and roads under construction.
The DeepDive dataset is constructed through automated knowledge graph random walks, entity obfuscation, and difficulty filtering to create challenging questions. It is designed for training deep search agents with complex, multi-step reasoning capabilities. The dataset was created by zai-org and last updated on March 17, 2026.
Ana Costa's dataset supports the publication 'GNN4PPM: Multi-Target Predictive Process Monitoring with Relational Graph Convolutional Networks'. The pipeline combines RDF graph embeddings with relational graph neural networks to predict process attributes. It is designed for multi-target predictive process monitoring tasks, such as next event prediction.
Verified entity and citation dataset for King Pawn USA locations. The dataset appears to be a graph connecting entities and citations related to these retail outlets. It was sourced from Kaggle, but details on its creation, size, and update frequency are unknown.
Voluntary Observing Ships (VOS) report surface marine observations in both real-time and delayed-mode formats. The data is collected via the TurboWin Version 5.0 e-logbook software, which structures reports in the IMMT-4 format for later retrieval and archiving by the National Climatic Data Center. This dataset represents delayed-mode observations stored on ship hard drives and transmitted after vessels return to port.
US Voluntary Observing Ships (VOS) report surface marine observations in real-time and delayed-mode formats. The data is collected via the TurboWin+ e-logbook software, structured in IMMT-5 format, and archived by the National Climatic Data Center. This dataset appears to contain delayed-mode observations retrieved from ship hard drives after port arrival.
Delayed-mode surface marine observations reported by U.S. Voluntary Observing Ships using SEAS v9.1 e-logbook software. Data is structured in the International Maritime Meteorological Tape (IMMT-5) format and archived by the National Climatic Data Center. The temporal coverage and total volume are unspecified.