A5066466375- Logic, Reasoning, and Knowledge
- Multi-Agent Systems and Negotiation
- Semantic Web and Ontologies
- Logic, programming, and type systems
- Advanced Database Systems and Queries
- Advanced Algebra and Logic
- AI-based Problem Solving and Planning
- Topic Modeling
- Bayesian Modeling and Causal Inference
- Data Management and Algorithms
- Constraint Satisfaction and Optimization
- Natural Language Processing Techniques
- Misinformation and Its Impacts
- Time Series Analysis and Forecasting
- Business Process Modeling and Analysis
- Sentiment Analysis and Opinion Mining
- Hate Speech and Cyberbullying Detection
- Formal Methods in Verification
- Scientific Computing and Data Management
- Explainable Artificial Intelligence (XAI)
- Adversarial Robustness in Machine Learning
- Biomedical Text Mining and Ontologies
- Service-Oriented Architecture and Web Services
- Text and Document Classification Technologies
- Gene Regulatory Network Analysis
University of Sheffield
2023-2025
Universidade Nova de Lisboa
2015-2024
Universidade Federal de São Carlos
2018
University of Lisbon
2004-2007
Iscte – Instituto Universitário de Lisboa
2004
In this paper we take a step towards using Argumentation in Social Networks and introduce Abstract Frameworks, an extension of Dung's Frameworks that incorporates social voting. We propose class semantics for these new prove some important non-trivial properties which are crucial their applicability Networks.
Adapters and Low-Rank Adaptation (LoRA) are parameter-efficient fine-tuning techniques designed to make the training of language models more efficient. Previous results demonstrated that these methods can even improve performance on some classification tasks. This paper complements existing research by investigating how influence computation costs compared full fine-tuning. We focus specifically multilingual text tasks (genre, framing, persuasion detection; with different input lengths,...
Abstract Credibility signals represent a wide range of heuristics typically used by journalists and fact-checkers to assess the veracity online content. Automating extraction credibility presents significant challenges due necessity training high-accuracy, signal-specific extractors, coupled with lack sufficiently large annotated datasets. This paper introduces Pastel ( P rompted we A k S upervision wi T h cr E dibility signa L s), weakly supervised approach that leverages language models...
Neural networks have been the key to solve a variety of different problems. However, neural network models are still regarded as black boxes, since they do not provide any human-interpretable evidence why output certain result. We address this issue by leveraging on ontologies and building small classifiers that map model's internal state concepts from an ontology, enabling generation symbolic justifications for models. Using image classification problem testing ground, we discuss how...
Hate speech and toxic comments are a common concern of social media platform users. Although these are, fortunately, the minority in platforms, they still capable causing harm. Therefore, identifying is an important task for studying preventing proliferation toxicity media. Previous work automatically detecting focus mainly English, with very few languages like Brazilian Portuguese. In this paper, we propose new large-scale dataset Portuguese tweets annotated as either or non-toxic different...
Logic programs under the stable model semantics, or answer-set programs, provide an expressive rule-based knowledge representation framework, featuring a formal, declarative and well-understood semantics. However, handling evolution of rule bases is still largely open problem. The AGM framework for belief change was shown to give inappropriate results when directly applied logic non-monotonic semantics such as models. approaches address this issue, developed so far, proposed update based on...
Deep neural network-based methods have recently enjoyed great popularity due to their effectiveness in solving difficult tasks. Requiring minimal human effort, they turned into an almost ubiquitous solution multiple domains. However, the size and complexity of typical network models' architectures, as well sub-symbolical nature representations generated by neuronal activations, networks are essentially opaque, making it nearly impossible explain humans reasoning behind decisions. We address...
Credibility signals represent a wide range of heuristics that are typically used by journalists and fact-checkers to assess the veracity online content. Automating task credibility signal extraction, however, is very challenging as it requires high-accuracy signal-specific extractors be trained, while there currently no sufficiently large datasets annotated with all signals. This paper investigates whether language models (LLMs) can prompted effectively set 18 produce weak labels for each...