A5085423462- Semantic Web and Ontologies
- Service-Oriented Architecture and Web Services
- Logic, Reasoning, and Knowledge
- Advanced Database Systems and Queries
- Distributed systems and fault tolerance
- Constraint Satisfaction and Optimization
- Data Management and Algorithms
- Optimization and Search Problems
- Fuzzy Logic and Control Systems
- Machine Learning and Algorithms
- Digital Games and Media
- Context-Aware Activity Recognition Systems
- Advanced Manufacturing and Logistics Optimization
- AI-based Problem Solving and Planning
- Explainable Artificial Intelligence (XAI)
- Logic, programming, and type systems
- Innovative Microfluidic and Catalytic Techniques Innovation
- Data Stream Mining Techniques
- Formal Methods in Verification
- Natural Language Processing Techniques
- Software Engineering Techniques and Practices
- Hate Speech and Cyberbullying Detection
- Advanced Control Systems Optimization
- Complex Network Analysis Techniques
- Smart Parking Systems Research
Linköping University
2014-2024
Utrecht University
2013
Stream reasoning can be defined as incremental over incrementally-available information. The formula progression procedure for Metric Temporal Logic (MTL) makes use of syntactic rewritings to incrementally evaluate formulas against states. Progression however assumes complete state information, which problematic when not all information is available or observed, such in qualitative spatial tasks robotics applications. In those cases, there may uncertainty out a set possible states represents...
DyKnow is a framework for stream reasoning aimed at robot applications that need to reason over wide and varying array of sensor data e.g. situation awareness. The extends the Robot Operating System (ROS). This paper presents architecture services behind DyKnow's run-time reconfiguration capabilities offers an analysis quantitative qualitative overhead. Runtime interesting advantages, such as fault recovery handling changes set computational information resources are available system....
We present FairX, an open-source Python-based benchmarking tool designed for the comprehensive analysis of models under umbrella fairness, utility, and eXplainability (XAI). FairX enables users to train bias-removal evaluate their fairness using a wide array metrics, data utility generate explanations model predictions, all within unified framework. Existing tools do not have way synthetic generated from fair generative models, also they support training either. In we add in collection our...
Qualitative spatio-temporal reasoning is an active research area in Artificial Intelligence. In many situations there a need to reason about intertemporal qualitative spatial relations, i.e. relations between regions at different time-points. However, these can never be explicitly observed since they are applications where the partly acquired by for example robotic system it therefore necessary infer relations. This problem has, best of our knowledge, not been studied before. The...
Modern robotic systems often consist of a growing set information-producing components that need to be appropriately connected for the system function properly. This is commonly done manually or through relatively simple scripts by specifying explicitly which connect. However, this process cumbersome and error-prone, does not scale well as more are introduced, lacks flexibility robustness at run-time. paper presents an algorithm setting up maintaining implicit subscriptions information its...
Data Fairness is a crucial topic due to the recent wide usage of AI powered applications. Most real-world data filled with human or machine biases and when those are being used train models, there chance that model will reflect bias in training data. Existing bias-mitigating generative methods based on GANs, Diffusion models need in-processing fairness objectives fail consider computational overhead while choosing computationally-heavy architectures, which may lead high demands, instability...
This work presents Fair4Free, a novel generative model to generate synthetic fair data using data-free distillation in the latent space. Fair4Free can on situation when is private or inaccessible. In our approach, we first train teacher create representation and then distil knowledge student (using smaller architecture). The process of distilling data-free, i.e. does not have access training dataset while distilling. After distillation, use distilled samples. Our extensive experiments show...
With Twitter and other microblogging services, users can easily express their opinion ideas in short text messages. A recent trend is that use the real-time property of these services to share opinions thoughts as events unfold on TV or real world. In context broadcasts, (over a mobile device, for example) referred second screen. This paper presents first characterization screen usage over playoffs major sports league. We present both temporal spatial analysis during end National Hockey...
A discussion on the role of ontologies and stream reasoning in Internet Things applications.
Imagine placing an online order on your way to the grocery store, then being able pick collected basket upon arrival or shortly after. Likewise, imagine any retail order, made ready for pickup in minutes instead of days. In realize such a low-latency automatic warehouse logistics system, solvers must be basket-aware. That is, it is more important that full (the basket) picked timely and fast, than single item quickly. Current state-of-the-art methods are not Nor they optimized positive...