The availability of vast amounts of visual data with diverse and fruitful features is a key factor for developing, verifying, and benchmarking advanced computer vision (CV) algorithms and architectures. Most visual datasets are created and curated for specific tasks or with limited data distribution for very specific fields of interest, and there is no unified approach to manage and access them across diverse sources, tasks, and taxonomies. This not only creates unnecessary overheads when building robust visual recognition systems, but also introduces biases into learning systems and limits the capabilities of data-centric AI. To address these problems, we propose the Vision Knowledge Graph (VisionKG), a novel resource that interlinks, organizes and manages visual datasets via knowledge graphs and Semantic Web technologies. It can serve as a unified framework facilitating simple access and querying of state-of-the-art visual datasets, regardless of their heterogeneous formats and taxonomies. One of the key differences between our approach and existing methods is that VisionKG is not only based on metadata but also utilizes a unified data schema and external knowledge bases to integrate, interlink, and align visual datasets. It enhances the enrichment of the semantic descriptions and interpretation at both image and instance levels and offers data retrieval and exploratory services via SPARQL and natural language empowered by Large Language Models (LLMs). VisionKG currently contains 617 million RDF triples that describe approximately 61 million entities, which can be accessed at https://vision.semkg.org and through APIs. With the integration of 37 datasets and four popular computer vision tasks, we demonstrate its usefulness across various scenarios when working with computer vision pipelines.

Current methods for embedding-based query answering over incomplete Knowledge Graphs (KGs) only focus on inductive reasoning, i.e., predicting answers by learning patterns from the data, and lack the complementary ability to do deductive reasoning, which requires the application of domain knowledge to infer further information. To address this shortcoming, we investigate the problem of incorporating ontologies into embedding-based query answering models by defining the task of embedding-based ontology-mediated query answering. We propose various integration strategies into prominent representatives of embedding models that involve (1) different ontology-driven data augmentation techniques and (2) adaptation of the loss function to enforce the ontology axioms. We design novel benchmarks for the considered task based on the LUBM and the NELL KGs and evaluate our methods on them. The achieved improvements in the setting that requires both inductive and deductive reasoning are from 20% to 55% in HITS@3.

A knowledge graph (KG) represents a set of entities and their relations. To explore the content of a large and complex KG, a convenient way is keyword-based querying. Traditional methods assign small weights to salient entities or relations, and answer an exploratory keyword query by computing a group Steiner tree (GST), which is a minimum-weight subgraph that connects all the keywords in the query. Recent studies have suggested improving the semantic cohesiveness of a query answer by minimizing the pairwise semantic distances between the entities in a subgraph, but it remains unclear how to efficiently compute such a semantically cohesive subgraph. In this paper, we formulate it as a quadratic group Steiner tree problem (QGSTP) by extending the classical minimum-weight GST problem which is NP-hard. We design two approximation algorithms for QGSTP and prove their approximation ratios. Furthermore, to improve their practical performance, we present heuristics including pruning and ranking strategies.

Knowledge graphs (KGs) are essential resources for many applications including Web search and question answering. As KGs are often automatically constructed, they may contain incorrect facts. Detecting them is a crucial, yet extremely expensive task. Prominent solutions detect and explain inconsistency in KGs with respect to accompanying ontologies that describe the KG domain of interest. Compared to machine learning methods they are more reliable and human-interpretable but scale poorly on large KGs. In this paper, we present a novel approach to dramatically speed up the process of detecting and explaining inconsistency in large KGs by exploiting KG abstractions that capture prominent data patterns. Though much smaller, KG abstractions preserve inconsistency and their explanations. Our experiments with large KGs (e.g., DBpedia and Yago) demonstrate the feasibility of our approach and show that it significantly outperforms the popular baseline.