246

A lack of demographic context in existing toxic speech datasets limits our understanding of how different age groups communicate online.In collaboration with funk, a German public service content network, this research introduces the first large-scale German dataset annotated for toxicity and enriched with platform-provided age estimates.The dataset includes 3,024 human-annotated and 30,024 LLM-annotated anonymized comments from Instagram, TikTok, and YouTube.To ensure relevance, comments were consolidated using predefined toxic keywords, resulting in 16.7% labeled as problematic.The annotation pipeline combined human expertise with state-of-theart language models, identifying key categories such as insults, disinformation, and criticism of broadcasting fees.The dataset reveals agebased differences in toxic speech patterns, with younger users favoring expressive language and older users more often engaging in disinformation and devaluation.This resource provides new opportunities for studying linguistic variation across demographics and supports the development of more equitable and age-aware content moderation systems.

This comprehensive survey examines the integration of knowledge-based approaches in autonomous driving systems, specifically focusing on trajectory prediction and planning. We extensively analyze various methodologies for incorporating domain knowledge, traffic rules, and common-sense reasoning into autonomous driving systems. The survey categorizes and analyzes approaches based on their knowledge representation and integration methods, ranging from purely symbolic to hybrid neuro-symbolic architectures. We examine recent developments in logic programming, foundation models for knowledge representation, reinforcement learning frame-works, and other emerging technologies incorporating domain knowledge. This work systematically reviews recent approaches, identifying key challenges, opportunities, and future research directions in knowledge-enhanced autonomous driving systems. Our analysis reveals emerging trends in the field, including the increasing importance of interpretable AI, the role of formal verification in safety-critical systems, and the potential of hybrid approaches that combine traditional knowledge representation with modern machine learning techniques.

Abstract Predicting stock price movements remains a major challenge in time series analysis. Despite extensive research on various machine learning techniques, few models have consistently achieved success in automated stock trading. One of the main challenges in stock price forecasting is that the optimal model changes over time due to market dynamics. This paper aims to predict stock prices using automated reinforcement learning algorithms and to analyse their efficiency compared with conventional methods. We automate DQN models and their variants, known for their adaptability, by continuously retraining them using recent data to capture market dynamics. We demonstrate that our dynamic models improve the accuracy of predicting the directions of various DAX stocks from 50.00% to approximately 60.00%, compared with conventional methods. Additionally, we conclude that dynamic models should be updated in response to shifts rather than at fixed intervals.

In today’s digital age, fake news has become a major problem with serious consequences, ranging from social unrest to political upheaval. New methods for detecting and mitigating fake news are required to address this issue. In this work, we propose incorporating contextual and network-aware features into the detection process. This involves analyzing not only the content of a news article but also the context in which it was shared and the network of users who shared it, i.e., the information diffusion. Thus, we propose GETAE, G raph Information E nhanced Deep Neural Ne T work Ensemble A rchitectur E for Fake News Detection, a novel ensemble architecture that uses textual content together with the social interactions to improve fake news detection. GETAE contains two Branches: the Text Branch and the Propagation Branch. The Text Branch combines Word and Transformer embeddings with a Deep Neural Network architecture based on feed-forward and bidirectional Recurrent Neural Networks ( [Bi]RNN ) to capture contextual features and generate a Text Content Embedding. This integrated approach allows for a more comprehensive understanding of the textual information. The Propagation Branch considers the information propagation within the graph network and proposes a Deep Learning architecture that employs Node Embeddings to create novel Propagation Embedding. GETAE’s Ensemble module combines the Text Content and Propagation Embeddings, to create a powerful and unique Propagation-Enhanced Content Embedding which is afterward used for classification. The experimental results obtained on two real-world publicly available datasets, i.e., Twitter15 and Twitter16, prove that this approach improves fake news detection and outperforms state-of-the-art models. • GETAE : graph-enhanced deep network for fake news detection using text and social data • Enhanced Text Content Embedding : combines complex lexical and syntactic features • New Propagation Embedding : captures information spread from a node to its network • Propagation-Enhanced Content Embedding : combines text, context, and propagation data • GETAE benchmark : validated via cross-validation, ablation, tuning on Twitter15 & 16

As global digitization continues to grow, technology becomes more affordable and easier to use, and social media platforms thrive, becoming the new means of spreading information and news. Communities are built around sharing and discussing current events. Within these communities, users are enabled to share their opinions about each event. Using Sentiment Analysis to understand the polarity of each message belonging to an event, as well as the entire event, can help to better understand the general and individual feelings of significant trends and the dynamics on online social networks. In this context, we propose a new ensemble architecture, EDSA-Ensemble (Event Detection Sentiment Analysis Ensemble), that uses Event Detection and Sentiment Analysis to improve the detection of the polarity for current events from Social Media. For Event Detection, we use techniques based on Information Diffusion taking into account both the time span and the topics. To detect the polarity of each event, we preprocess the text and employ several Machine and Deep Learning models to create an ensemble model. The preprocessing step includes several word representation models: raw frequency,$TFIDF$, Word2Vec, and Transformers. The proposed EDSA-Ensemble architecture improves the event sentiment classification over the individual Machine and Deep Learning models.

This study investigates the performance of the DI-NOv2 pre-trained model within Federated Learning (FL) environments, focusing on its application to segmentation tasks across diverse domains. While DINOv2 has demonstrated high efficacy in centralized training scenarios, its capabilities under FL conditions—where data privacy and security are paramount—remain underexplored. Utilizing data sets spanning industrial, medical, and automotive sectors, we evaluated DINOv2’s accuracy and generalization in decentralized settings. Our findings reveal that federated DINOv2 performs comparably to centralized models, effectively segmenting objects despite the decentralized and heterogeneous nature of the data. However, inherent biases in the pre-trained model posed challenges, affecting performance across different domains. These results highlight the need for domain-specific fine-tuning and bias mitigation strategies to enhance the robustness of pre-trained models in FL contexts. Future work should address these challenges to maximize the potential of FL in privacy-sensitive applications, ensuring high performance while maintaining data confidentiality.

Autonomous agents often face the challenge of interpreting uncertain natural language instructions for planning tasks. Representing these instructions as Linear Temporal Logic (LTL) enables planners to synthesize actionable plans. We introduce CoT-TL, a data-efficient in-context learning framework for translating natural language specifications into LTL representations. CoT-TL addresses the limitations of large language models, which typically rely on extensive fine-tuning data, by extending chain-of-thought reasoning and semantic roles to align with the requirements of formal logic creation. This approach enhances the transparency and rationale behind LTL generation, fostering user trust. CoT-TL achieves state-of-the-art accuracy across three diverse datasets in low-data scenarios, outperforming existing methods without fine-tuning or intermediate translations. To improve reliability and minimize hallucinations, we incorporate model checking to validate the syntax of the generated LTL output. We further demonstrate CoT-TL’s effectiveness through ablation studies and evaluations on unseen LTL structures and formulas in a new dataset. Finally, we validate CoT-TL’s practicality by integrating it into a QuadCopter for multi-step drone planning based on natural language instructions.

Traffic rules formalization is crucial for verifying the compliance and safety of autonomous vehicles (AVs). However, manual translation of natural language traffic rules as formal specification requires domain knowledge and logic expertise, which limits its adaptation. This paper introduces TR2MTL, a framework that employs large language models (LLMs) to automatically translate traffic rules (TR) into metric temporal logic (MTL). It is envisioned as a human-in-loop system for AV rule formalization. It utilizes a chain-of-thought in-context learning approach to guide the LLM in step-by-step translation and generating valid and grammatically correct MTL formulas. It can be extended to various forms of temporal logic and rules. We evaluated the framework on a challenging dataset of traffic rules we created from various sources and compared it against LLMs using different in-context learning methods. Results show that TR2MTL is domain-agnostic, achieving high accuracy and generalization capability even with a small dataset. Moreover, the method effectively predicts formulas with varying degrees of logical and semantic structure in unstructured traffic rules.

Real-time social media data can provide useful information on evolving hazards. Alongside traditional methods of disaster detection, the integration of social media data can considerably enhance disaster management. In this paper, we investigate the problem of detecting geolocation-content communities on Twitter and propose a novel distributed system that provides in near real-time information on hazard-related events and their evolution. We show that content-based community analysis can lead to better and faster dissemination of hazard-related reports than using only traditional methods, such as satellite or airborne sensing platforms. Our distributed disaster reporting system analyzes the social relationship among worldwide geolocated tweets and applies topic modeling to group tweets by topics. Considering for each tweet the following information: user, timestamp, geolocation, retweets, and replies, we create a publisher-subscriber distribution model for topics. We use content similarity and the proximity of nodes to create a new model for geolocation-content based communities. Users can subscribe to different topics in specific geographical areas or worldwide and receive real-time reports regarding these topics. As misinformation can lead to increased damage if propagated in hazards-related tweets, we propose a new deep learning model to detect fake news. The misinformed tweets are then removed from display. We also show empirically the scalability capabilities of the proposed system.

With the increase of digital communication, a surge in online hate speech can be witnessed. Recent studies have concentrated on automated supervised detection of hate speech. However, there remains limited understanding of an effective strategy for identifying multilingual hate speech in social media posts. This study introduces an innovate experimental design for multilingual hate speech detection. It compares different approaches to automatically detect multilingual hate speech through a series of experiments and creates a classification algorithm for hate speech in German, Italian and Spanish text-based social media content. The study creates monolingual, multilingual, and translated datasets specific to the language triplet. Subsequently, the research explores suitable models for multilingual hate speech detection, evaluating a total of seven transformer-based models along with corresponding SVM models on the constructed datasets. The findings indicate that all chosen transformer-based models outperform the baseline SVM models. The research highlights the superiority of a multilingual approach, utilizing XLM-RoBERTa as a classifier model, over monolingual, multilingual, and translation-based approaches. Furthermore, the study demonstrates that translation-based methods in connection to the model DistillBERT can serve as viable alternatives to the multilingual XLM-RoBERTa approach, particularly in scenarios where computational resources are restricted and processing speed is of importance.

Information extraction tasks are particularly challenging in specific contexts such as the legal domain. In this paper, Named Entity Recognition is used to make legal texts more accessible to domain experts and laymen. This paper focuses on extracting law references and citations of court decisions, which occur in various syntactic formats. To investigate this task a reference data set is constructed from a large collection of German court decisions and different NER-techniques are compared. Pattern matching, probabilistic sequence labeling (CRF), Deep Learning (BiLSTM) and transfer learning using a pretrained language model (BERT) are applied to extract references to laws and court decisions. The results show that the BERT based approach achieves F1 scores around 0.98 for both tasks and outperforms methods from prior work, which achieve F1 scores of 0.89 (CRF for law references) respectively 0.82 (CRF for court decisions) on the same data set.

Topic modeling is a well-known task in unsupervised machine learning, where clustering algorithms are used to find latent topics. Several algorithms are presented in the literature, but the best known of them suffer from the drawback of requiring a lot of hyperparameter tuning to achieve good results. Especially, the number of latent topics or clusters (k) needs to be known in advance. In view of this situation, this paper analyses objective functions that help to evaluate the models in order to determine optimal hyperparameters. An empirical qualitative study was conducted using the NMF algorithm on different datasets to experimentally determine numerical properties of topic models which indicate an optimal k. Based on this study, we propose objective functions to select optimal topic models and discuss their results on different datasets.

Aspect-Oriented Programming (AOP) is a technology for the decomposition of software systems based on cross-cutting concerns. As shown in our previous work, cross-cutting concerns are also present in ontologies, and Aspect-Oriented Ontology Development (AOOD) can be used for flexible and dynamic ontology modularization based on functional and non-functional requirements. When ontologies are used in applications, application and ontology-related requirements often coincide. In this paper, we show that aspects in ontologies can be expressed as software aspects and directly referred to from software code using the well-known AspectJ language and Java annotations. We present an extension of the well-known OWL API with aspect-oriented means that allow transparent access to and manipulation of ontology modules that are based on requirements.

In this paper, we propose a dynamic and flexible approach to ontology modularization inspired by the aspect-oriented programming paradigm in software development. Aspect-oriented programming provides formalisms for encapsulating a software system's functionality in self-contained modules, where each module represents a single requirement, along with meta-information on how to recombine the modules at runtime. The recombination requires second-order reasoning, but it can be shown that it is sufficient to interpret the second-order constructs under a Henkin semantics, reducing the problem to first-order logic. We demonstrate the applicability of the approach to the problem of modular ontologies by presenting a proof-of concept implementation for aspect-oriented OWL 2 ontologies. Furthermore, we show that the approach can be used as a substitute for metamodeling and may be helpful in preventing expensive refactoring operations in certain ontology integration scenarios.

In this paper we present the motivation, use cases, design principles, abstract syntax, and initial core of LegalRuleML. The LegalRuleML-core is sufficiently rich for expressing legal sources, time, defeasibility, and deontic operators. An example is provided. LegalRuleMLis compared to related work.

Usage of background knowledge about events and their relations to other concepts in the application domain, can improve the quality of event processing. In this paper, we describe a system for knowledge-based event detection of complex stock market events based on available background knowledge about stock market companies. Our system profits from data fusion of live event stream and background knowledge about companies which is stored in a knowledge base. Users of our system can express their queries in a rule language which provides functionalities to specify semantic queries about companies in the SPARQL query language for querying the external knowledge base and combine it with event data stream. Background makes it possible to detect stock market events based on companies attributes and not only based on syntactic processing of stock price and volume.

In the field of software engineering component-based development and appropriate documentation are established methods to support reuse. While modular development is tackled in various work regarding ontology engineering, it is an open problem how documentation of ontologies should be created. After analyzing existing ontology documentations we identified grouping concepts as a very helpful technique to simplify the understandability and thus improve the reusability of ontologies. In this paper, we present a technique to group concepts for ontology documentation by applying community detection algorithms on the graph structure of ontologies. Using the manually created concept groups from existing documentations as reference we demonstrate that this technique is able to create appropriate concept groups automatically.

Most patent applications almost always face non-obvious/inventive step rejections during the examination stage. The rejections based on non-obviousness/inventive step are increasing substantially each year. In this paper, we propose a mathematical approach called the FSTP Test for determining a non-obviousness indication. The FSTP Test allows an inventor to identify and rework certain aspects of his invention before filing a patent application, which might have been considered as obvious at a later stage.

Abstract As Semantic Web technologies mature and new releases of key elements, such as SPARQL 1.1 and OWL 2.0, become available, the Life Sciences continue to push the boundaries of these technologies with ever more sophisticated tools and applications. Unsurprisingly, therefore, interest in the SWAT4LS (Semantic Web Applications and Tools for the Life Sciences) activities have remained high, as was evident during the third international SWAT4LS workshop held in Berlin in December 2010. Contributors to this workshop were invited to submit extended versions of their papers, the best of which are now made available in the special supplement of BMC Bioinformatics. The papers reflect the wide range of work in this area, covering the storage and querying of Life Sciences data in RDF triple stores, tools for the development of biomedical ontologies and the semantics-based integration of Life Sciences as well as clinicial data.

The eight papers in this special section focus on the state-of-the-art approaches, solutions, and applications in the area of rule representation, reasoning, and interchange in the context of distributed, (partially) open, heterogeneous environments, such as the semantic Web, intelligent multiagent systems, event-driven architectures. and service-oriented computing.

Background: Semantically-enriched browsing has enhanced the browsing experience by providing contextualised dynamically generated Web content, and quicker access to searched-for information. However, adoption of Semantic Web technologies is limited and user perception from the non-IT domain sceptical. Furthermore, little attention has been given to evaluating semantic browsers with real users to demonstrate the enhancements and obtain valuable feedback. The Sealife project investigates semantic browsing and its application to the life science domain. Sealife's main objective is to develop the notion of context-based information integration by extending three existing Semantic Web browsers (SWBs) to link the existing Web to the eScience infrastructure. Methods: This paper describes a user-centred evaluation framework that was developed to evaluate the Sealife SWBs that elicited feedback on users' perceptions on ease of use and information findability. Three sources of data: i) web server logs; ii) user questionnaires; and iii) semi-structured interviews were analysed and comparisons made between each browser and a control system. Results: It was found that the evaluation framework used successfully elicited users' perceptions of the three distinct SWBs. The results indicate that the browser with the most mature and polished interface was rated higher for usability, and semantic links were used by the users of all three browsers. Conclusion: Confirmation or contradiction of our original hypotheses with relation to SWBs is detailed along with observations of implementation issues. © 2009 Oliver et al; licensee BioMed Central Ltd.

BACKGROUND: As interest in adopting the Semantic Web in the biomedical domain continues to grow, Semantic Web technology has been evolving and maturing. A variety of technological approaches including triplestore technologies, SPARQL endpoints, Linked Data, and Vocabulary of Interlinked Datasets have emerged in recent years. In addition to the data warehouse construction, these technological approaches can be used to support dynamic query federation. As a community effort, the BioRDF task force, within the Semantic Web for Health Care and Life Sciences Interest Group, is exploring how these emerging approaches can be utilized to execute distributed queries across different neuroscience data sources. METHODS AND RESULTS: We have created two health care and life science knowledge bases. We have explored a variety of Semantic Web approaches to describe, map, and dynamically query multiple datasets. We have demonstrated several federation approaches that integrate diverse types of information about neurons and receptors that play an important role in basic, clinical, and translational neuroscience research. Particularly, we have created a prototype receptor explorer which uses OWL mappings to provide an integrated list of receptors and executes individual queries against different SPARQL endpoints. We have also employed the AIDA Toolkit, which is directed at groups of knowledge workers who cooperatively search, annotate, interpret, and enrich large collections of heterogeneous documents from diverse locations. We have explored a tool called "FeDeRate", which enables a global SPARQL query to be decomposed into subqueries against the remote databases offering either SPARQL or SQL query interfaces. Finally, we have explored how to use the vocabulary of interlinked Datasets (voiD) to create metadata for describing datasets exposed as Linked Data URIs or SPARQL endpoints. CONCLUSION: We have demonstrated the use of a set of novel and state-of-the-art Semantic Web technologies in support of a neuroscience query federation scenario. We have identified both the strengths and weaknesses of these technologies. While Semantic Web offers a global data model including the use of Uniform Resource Identifiers (URI's), the proliferation of semantically-equivalent URI's hinders large scale data integration. Our work helps direct research and tool development, which will be of benefit to this community.