Prof. Dr.

Hauswirth, Manfred

Filters

1
4 2
3 3
Reset filters

Settings
Now showing 1 - 6 of 6
No Thumbnail Available
Publication

VisionKG: Unleashing the Power of Visual Datasets via Knowledge Graph

2023-09 , Yuan, Jicheng , Le-Tuan, Anh , Nguyen-Duc, Manh , Tran, Trung-Kien , Hauswirth, Manfred , Phuoc, Danh Le

The availability of vast amounts of visual data with heterogeneous features is a key factor for developing, testing, and benchmarking of new computer vision (CV) algorithms and architectures. Most visual datasets are created and curated for specific tasks or with limited image data distribution for very specific situations, 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 ours is knowledge-based rather than metadatabased. It enhances the enrichment of the semantics at both image and instance levels and offers various data retrieval and exploratory services via SPARQL. VisionKG currently contains 519 million RDF triples that describe approximately 40 million entities, and are accessible at https://vision.semkg.org and through APIs. With the integration of 30 datasets and four popular CV tasks, we demonstrate its usefulness across various scenarios when working with CV pipelines.

View
No Thumbnail Available
Publication

Towards building live open scientific knowledge graphs

2022 , Le-Tuan, Anh , Franzreb, Carlos , Phuoc, Danh Le , Schimmler, Sonja , Hauswirth, Manfred

Due to the large number and heterogeneity of data sources, it becomes increasingly difficult to follow the research output and the scientific discourse. For example, a publication listed on DBLP may be discussed on Twitter and its underlying data set may be used in a different paper published on arXiv. The scientific discourse this publication is involved in is divided among not integrated systems, and for researchers it might be very hard to follow all discourses a publication or data set may be involved in. Also, many of these data sources-DBLP, arXiv, or Twitter, to name a few-are often updated in real-time. These systems are not integrated (silos), and there is no system for users to query the content/data actively or, what would be even more beneficial, in a publish/subscribe fashion, i.e., a system would actively notify researchers of work interesting to them when such work or discussions become available. In this position paper, we introduce our concept of a live open knowledge graph which can integrate an extensible set of existing or new data sources in a streaming fashion, continuously fetching data from these heterogeneous sources, and interlinking and enriching it on-the-fly. Users can subscribe to continuously query the content/data of their interest and get notified when new content/data becomes available. We also highlight open challenges in realizing a system enabling this concept at scale.

View
No Thumbnail Available
Publication

Towards a decentralized data hub and query system for federated dynamic data spaces

2023 , Phuoc, Danh Le , Schimmler, Sonja , Le-Tuan, Anh , Kuehn, Uwe A. , Hauswirth, Manfred

This position paper proposes a hybrid architecture for secure and efficient data sharing and processing across dynamic data spaces. On the one hand, current centralized approaches are plagued by issues such as lack of privacy and control for users, high costs, and bad performance, making these approaches unsuitable for the decentralized data spaces prevalent in Europe and various industries (decentralized on the conceptual and physical levels while centralized in the underlying implementation). On the other hand, decentralized systems face challenges with limited knowledge of/control over the global system, fair resource utilization, and data provenance. Our proposed Semantic Data Ledger (SDL) approach combines the advantages of both architectures to overcome their limitations. SDL allows users to choose the best combination of centralized and decentralized features, providing a decentralized infrastructure for the publication of structured data with machine-readable semantics. It supports expressive structured queries, secure data sharing, and payment mechanisms based on an underlying autonomous ledger, enabling the implementation of economic models and fair-use strategies.

View
No Thumbnail Available
Publication

CQELS 2.0: Towards a unified framework for semantic stream fusion

2022 , Le-Tuan, Anh , Nguyen-Duc, Manh , Le, Chien-Quang , Tran, Trung-Kien , Hauswirth, Manfred , Eiter, Thomas , Phuoc, Danh Le

We present CQELS 2.0, the second version of Continuous Query Evaluation over Linked Streams. CQELS 2.0 is a platform-agnostic federated execution framework towards semantic stream fusion. In this version, we introduce a novel neuralsymbolic stream reasoning component that enables specifying deep neural network (DNN) based data fusion pipelines via logic rules with learnable probabilistic degrees as weights. As a platform-agnostic framework, CQELS 2.0 can be implemented for devices with different hardware architectures (from embedded devices to cloud infrastructures). Moreover, this version also includes an adaptive federator that allows CQELS instances on different nodes in a network to coordinate their resources to distribute processing pipelines by delegating partial workloads to their peers via subscribing continuous queries.

View
No Thumbnail Available
Publication

SemRob: Towards semantic stream reasoning for robotic operating systems

2022 , Nguyen-Duc, Manh , Le-Tuan, Anh , Hauswirth, Manfred , Bowden, David , Phuoc, Danh Le

Stream processing and reasoning is getting considerable attention in various application domains such as IoT, Industry IoT and Smart Cities. In parallel, reasoning and knowledge-based features have attracted research into many areas of robotics, such as robotic mapping, perception and interaction. To this end, the Semantic Stream Reasoning (SSR) framework can unify the representations of symbolic/semantic streams with deep neural networks, to integrate high-dimensional data streams, such as video streams and LiDAR point clouds, with traditional graph or relational stream data. As such, this positioning and system paper will outline our approach to build a platform to facilitate semantic stream reasoning capabilities on a robotic operating system called SemRob.

View
No Thumbnail Available
Publication

VisionKG: Towards a unified vision knowledge graph

2021 , Le-Tuan, Anh , Tran, Trung-Kien , Nguyen-Duc, Manh , Yuan, Jicheng , Hauswirth, Manfred , Phuoc, Danh Le

Computer Vision (CV) has recently achieved signi_cant im-provements, thanks to the evolution of deep learning. Along with ad-vanced architectures and optimisations of deep neural networks, CV data for (cross-datasets) training, validating, and testing contributes greatly to the performance of CV models. Many CV datasets have been created for different tasks, but they are available in heterogeneous data formats and semantic representations. Therefore, it is challenging when one needs to combine different datasets either for training or testing purposes. This paper proposes a unified framework using the Semantic Web technology that provides a novel way to interlink and integrate labelled data across different data sources. We demonstrate its advantages via various sce-narios with the system framework accessible both online and via APIs.4.

View