In recent years, Federated Learning (FL) has emerged as a privacy-preserving paradigm for collaborative model training in IoT systems, enabling clients to learn a global model for tasks like classification, prediction, or anomaly detection in IoT environments without sharing raw data. However, traditional centralized FL architectures face bottlenecks, single points of failure, and struggle with non-IID data. These limitations hinder effective Collective Intelligence in large-scale IoT systems where numerous devices operate across diverse and dynamic environments. Existing clustered FL approaches often retain centralization or overlook how the spatial distribution inherent in IoT deployments directly influences data heterogeneity, challenging both the integration of spatially correlated devices and the establishment of intelligence distributed across the entire system. Creating such intelligence demands both decentralized architectures for scalability and effective integration of devices with similar data distributions. For these reasons, this article introduces Proximity-Aware Self-Federated Learning (PSFL), a novel decentralized approach embodying collective intelligence principles. PSFL leverages field-based coordination to enable IoT devices to form self-federations, dynamically clustered groups that train specialized models based on both spatial proximity and local model characteristics. These self-federations reflect underlying data distributions, creating a distributed ecosystem of specialized models across the network. This approach overcomes global model limitations in non-IID settings through specialized federations based on local data distributions, enhancing performance while maintaining decentralization. We evaluate our approach using the Extended MNIST and CIFAR-100 datasets against state-of-the-art baselines, demonstrating its effectiveness in forming coherent, localized models under non-IID conditions.

We present mstlo (mistletoe), a Rust library for high-performance online monitoring of signal temporal logic (STL), with Python bindings. The library provides: (i) a unified interface for multiple STL semantics, including Robust Satisfaction Intervals (RoSI) and Boolean evaluation with early verdicts; (ii) an incremental monitoring algorithm based on bottom-up dynamic programming with per-operator caching and streaming extremum computation for temporal operators; and (iii) an embedded STL domain-specific language for both Rust and Python implementations, with procedural macros in Rust for static syntax checking. Benchmarks show scalability and performance improvements over state-of-the-art tools, especially for formulas with large temporal depth and deep nesting.

In the last years, Federated learning (FL) has become a popular solution to train machine learning models in domains with high privacy concerns. However, FL scalability and performance face significant challenges in real-world deployments where data across devices are non-independently and identically distributed (non-IID). The heterogeneity in data distribution frequently arises from spatial distribution of devices, leading to degraded model performance in the absence of proper handling. Additionally, FL typical reliance on centralized architectures introduces bottlenecks and single-point-of-failure risks, particularly problematic at scale or in dynamic environments. To close this gap, we propose Field-Based Federated Learning (FBFL), a novel approach leveraging macroprogramming and field coordination to address these limitations through: (i) distributed spatial-based leader election for personalization to mitigate non-IID data challenges; and (ii) construction of a self-organizing, hierarchical architecture using advanced macroprogramming patterns. Moreover, FBFL not only overcomes the aforementioned limitations, but also enables the development of more specialized models tailored to the specific data distribution in each subregion. This paper formalizes FBFL and evaluates it extensively using MNIST, FashionMNIST, and Extended MNIST datasets. We demonstrate that, when operating under IID data conditions, FBFL performs comparably to the widely-used FedAvg algorithm. Furthermore, in challenging non-IID scenarios, FBFL not only outperforms FedAvg but also surpasses other state-of-the-art methods, namely FedProx and Scaffold, which have been specifically designed to address non-IID data distributions. Additionally, we showcase the resilience of FBFL's self-organizing hierarchical architecture against server failures.

Coordination of view coverage via privacy-aware smart cameras is key to a more socially responsible urban intelligence. Rather than maximizing view coverage at any cost or over relying on expensive cryptographic techniques, we address how cameras can coordinate to legitimately monitor public spaces while excluding privacy-sensitive regions by design. This article proposes a decentralized framework in which interactive smart cameras coordinate to autonomously select their orientation via collective learning, while eliminating privacy violations via soft and hard constraint satisfaction. The approach scales to hundreds up to thousands of cameras without any centralized control. Experimental evidence shows 18.42

OBJECTIVES:Federated learning (FL) may mitigate privacy constraints, heterogeneous data quality, and inconsistent labelling in dental artificial intelligence (AI). FL was compared with centralized (CL) and local learning (LL) for tooth segmentation in panoramic radiographs across multiple data corruption scenarios. METHODS:An Attention U-Net was trained on 2066 radiographs from 6 institutions across 4 settings: baseline (unaltered data), label manipulation (dilated/missing annotations), image-quality manipulation (additive Gaussian noise), and exclusion of one faulty client with corrupted data. FL was implemented via the Flower AI framework. Per-client training and validation loss trajectories were monitored for anomaly detection and a set of metrics (Dice, IoU, HD, HD95, and ASSD) were evaluated on a hold-out test set. From these metrics significance results were reported through Wilcoxon signed-rank test. CL and LL served as comparators. RESULTS:Baseline: FL achieved a median Dice of 0.949 (ASSD: 1.332), slightly better than CL at 0.947 (ASSD: 1.371) and LL at 0.936-0.940 (ASSD: 1.519-1.698). Label manipulation: FL maintained the best median Dice score at 0.949 (ASSD: 1.465) versus CL's 0.942 (ASSD: 1.757) and LL's 0.930-0.940 (ASSD: 1.519-2.115). Similar performance was observed when 2 faulty clients were introduced. Image noise: FL led with a Dice at 0.949 (ASSD: 1.311); CL had a Dice of 0.948 (ASSD: 1.361); LL ranged from 0.932 to 0.940 (ASSD: 1.519-1.774). Similar performance was observed when 2 faulty clients were introduced, with CL performing slightly better than FL. Faulty client exclusion: FL showed a Dice of 0.948 (ASSD: 1.331) better than CL's 0.946 (ASSD: 1.393). Loss curve monitoring reliably flagged the corrupted site. CONCLUSIONS:FL matches or exceeds CL and outperforms LL across corruption scenarios while preserving privacy. Per-client loss trajectories provide an effective anomaly-detection mechanism and support FL as a practical, privacy-preserving approach for scalable clinical AI development.

Multi-agents systems are becoming rather pervasive and are applicable in several domains ranging from logistics, healthcare to manufacturing etc. As individual agents work towards fulfilling their own goals, possibly collaborating with each other if needed, e.g., an agent becomes incapacitated, thus needing help to complete its tasks, deciding which agents to trust and interact with is of outmost importance. Leveraging trust in multi-agent systems can lead to an improvement in performance, as agents do not waste resources with malicious entities. In this paper, we propose a model relevant for a job distribution scenario in which multiple agents collaboratively solve a job by dividing it into multiple tasks, thereafter assigning these tasks based on a calculated value for trust. Furthermore, we provide a simulation setup that can be used to test different evaluation and rating functions, across different numbers of benign and malicious agents, and other configurable parameters. Our results show that with the proposed model agents can in general correctly classify benign and malicious agents, with the results being stable across four different rating functions.

Recent advances in large-scale video models have significantly improved video understanding across domains such as surveillance, healthcare, and entertainment. However, these models also amplify privacy risks by encoding sensitive attributes, including facial identity, race, and gender. While image anonymization has been extensively studied, video anonymization remains relatively underexplored, even though modern video models can leverage spatiotemporal motion patterns as biometric identifiers. To address this challenge, we propose a novel attention-driven spatiotemporal video anonymization framework based on systematic disentanglement of utility and privacy features. Our key insight is that attention mechanisms in Vision Transformers (ViTs) can be explicitly structured to separate action-relevant information from privacy-sensitive content. Building on this insight, we introduce two task-specific classification tokens, an action CLS token and a privacy CLS token, that learn complementary representations within a shared Transformer backbone. We contrast their attention distributions to compute a utility-privacy score for each spatiotemporal tubelet, and keep the top-k tubelets with the highest scores. This selectively prunes tubelets dominated by privacy cues while preserving those most critical for action recognition. Extensive experiments demonstrate that our approach maintains action recognition performance comparable to models trained on raw videos, while substantially reducing privacy leakage. These results indicate that attention-driven spatiotemporal pruning offers an effective and principled solution for privacy-preserving video analytics.

Collective Adaptive Systems (CAS) increasingly rely on machine learning to let each node learn from locally sensed data, aligning its behavior with the surrounding environment. Scaling this intelligence, however, raises fundamental challenges: sensed data is often privacy-sensitive, preventing centralized collection; nodes are mobile, traversing regions where nearby nodes perceive similar phenomena while distant ones observe radically different conditions, creating natural spatial clusters; and these distributions evolve over time due to mobility, introducing temporal drift that makes local models progressively stale. These dynamics arise across domains - vehicular sensing, drone-based monitoring, smartphone crowdsensing - yet the interplay of privacy, spatial heterogeneity, and temporal drift severely undermines conventional learning strategies. Therefore, we propose C2FL, a fully distributed Federated Learning (FL) approach where nodes self-organize into learning groups through spatial clustering, reflecting the geographic structure of the environment. To counteract temporal drift, each node combines experience replay with a dwell-time-aware adaptive averaging step, progressively incorporating the regional consensus as it remains longer within the same area, while preserving previously acquired knowledge under evolving distributions. We evaluate our approach on synthetic experiments that systematically reproduce spatial and temporal shifts, showing that standard federated strategies degrade significantly under these conditions and that our method restores robust collective adaptation.

Self-adaptive robotic systems operate autonomously in dynamic and uncertain environments, requiring robust real-time monitoring and adaptive behaviour. Unlike traditional robotic software with predefined logic, self-adaptive robots exploit artificial intelligence (AI), machine learning, and model-driven engineering to adapt continuously to changing conditions, thereby ensuring reliability, safety, and optimal performance. This paper presents a research agenda for software engineering in self-adaptive robotics, structured along two dimensions. The first concerns the software engineering lifecycle, requirements, design, development, testing, and operations, tailored to the challenges of self-adaptive robotics. The second focuses on enabling technologies such as digital twins and AI-driven adaptation, which support runtime monitoring, fault detection, and automated decision-making. We identify open challenges, including verifying adaptive behaviours under uncertainty, balancing trade-offs between adaptability, performance, and safety, and integrating self-adaptation frameworks like MAPE K/MAPLE-K. By consolidating these challenges into a roadmap toward 2030, this work contributes to the foundations of trustworthy and efficient self-adaptive robotic systems capable of meeting the complexities of real-world deployment.

Federated Learning (FL) has emerged as a powerful paradigm for training machine learning models across distributed data sources while preserving data locality. However, the privacy of local data is always a pivotal concern and has received a lot of attention in recent research on the FL regime. Moreover, the lack of domain heterogeneity and client-specific segregation in the benchmarks remains a critical bottleneck for rigorous evaluation. In this paper, we introduce ModelNet, a novel image classification dataset constructed from the embeddings extracted from a pre-trained ResNet50 model. First, we modify the CIFAR100 dataset into three client-specific variants, considering three domain heterogeneities (homogeneous, heterogeneous, and random). Subsequently, we train each client-specific subset of all three variants on the pre-trained ResNet50 model to save model parameters. In addition to multi-domain image data, we propose a new hypothesis to define the FL algorithm that can access the anonymized model parameters to preserve the local privacy in a more effective manner compared to existing ones. ModelNet is designed to simulate realistic FL settings by incorporating non-IID data distributions and client diversity design principles in the mainframe for both conventional and futuristic graph-driven FL algorithms. The three variants are ModelNet-S, ModelNet-D, and ModelNet-R, which are based on homogeneous, heterogeneous, and random data settings, respectively. To the best of our knowledge, we are the first to propose a cross-environment client-specific FL dataset along with the graph-based variant. Extensive experiments based on domain shifts and aggregation strategies show the effectiveness of the above variants, making it a practical benchmark for classical and graph-based FL research. The dataset and related code are available online.

Digital twin (DT) platforms are limited by resource requirements. While scaling such platforms would be possible, the associated cost could be rather expensive. Furthermore, the number of DT-enabled systems is increasing, ranging from individual DTs to fleets of DTs and composite DTs. To handle the complexity, we propose in this paper four modes of operation for DTs, including individual DT execution, fleets, composites, and multi-fidelity DTs. For each, we present a resource profile in order to characterize their operation. We showcase our framework and the resource profiling through a case-study, covering the case of individual DT execution.

Large Language Models (LLMs) have demonstrated remarkable capabilities in generating human-like text, yet their applicability to dialogue systems in computer games remains limited. This limitation arises from their substantial hardware requirements, latency constraints, and the necessity to maintain clearly defined knowledge boundaries within a game setting. In this paper, we propose a modular NPC dialogue system that leverages Small Language Models (SLMs), fine-tuned to encode specific NPC personas and integrated with runtime-swappable memory modules. These memory modules preserve character-specific conversational context and world knowledge, enabling expressive interactions and long-term memory without retraining or model reloading during gameplay. We comprehensively evaluate our system using three open-source SLMs: DistilGPT-2, TinyLlama-1.1B-Chat, and Mistral-7B-Instruct, trained on synthetic persona-aligned data and benchmarked on consumer-grade hardware. While our approach is motivated by applications in gaming, its modular design and persona-driven memory architecture hold significant potential for broader adoption in domains requiring expressive, scalable, and memory-rich conversational agents, such as virtual assistants, customer support bots, or interactive educational systems.

Federated Learning (FL) has emerged as a powerful paradigm for training machine learning models across distributed data sources while preserving data locality. However, the privacy of local data is always a pivotal concern and has received a lot of attention in recent research on the FL regime. Moreover, the lack of domain heterogeneity and client-specific segregation in the benchmarks remains a critical bottleneck for rigorous evaluation. In this paper, we introduce ModelNet, a novel image classification dataset constructed from the embeddings extracted from a pre-trained ResNet50 model. First, we modify the CI-F AR100 dataset into three client-specific variants, considering three domain heterogeneities (homogeneous, heterogeneous, and random). Subsequently, we train each client-specific subset of all three variants on the pre-trained ResNet50 model to save model parameters. In addition to multi-domain image data, we propose a new hypothesis to define the FL algorithm that can access the anonymized model parameters to preserve the local privacy in a more effective manner compared to existing ones. ModelNet is designed to simulate realistic FL settings by incorporating non-IID data distributions and client diversity design principles in the mainframe for both conventional and futuristic graph-driven FL algorithms. The three variants are ModelNet-S, ModelNet-D, and ModelNet-R, which are based on homogeneous, heterogeneous, and random data settings, respectively. To the best of our knowledge, we are the first to propose a cross-environment client-specific FL dataset along with the graph-based variant. Extensive experiments based on domain shifts and aggregation strategies show the effectiveness of the above variants, making it a practical benchmark for classical and graph-based FL research. The dataset and related code are available here11https://github.com/rayabhisek123/ModelNet

Systems that adapt to their environment or change based on new requirements pose challenges for runtime verification. Complexity is increased when the system needs to retain its internal state and continue monitoring while also updating properties or adding new ones during runtime. In this work, we propose DynSRV, a Stream Runtime Verification language that allows for dynamic updates of properties. A core benefit of this language is its capability to update properties at runtime without requiring a restart of the monitor, maintaining the internal state of the remaining properties. We formalise the semantics of our core primitives and demonstrate design patterns for allowing adaptations under certain constraints. Finally, we present an implementation of DynSRV and describe three memory strategies that balance memory usage and the ability to resolve dynamically added properties depending on historical data.

The MADTECC 2024 chairs welcome you to the 1st Workshop on Medical Applications with Digital Twins and Edge-cloud Continuum

Causal modeling and cause-detection (abduction) are critical challenges for understanding and managing complex adaptive systems (CAS), including self-improving and self-integrating systems (SISSY). While the vast research literature on causality addresses many important issues, it mostly lacks specific methods for dealing with CAS-related issues - e.g. runtime changes and uncertainty; integrating new types of variables on-the-fly; nonlinear phenomena and large system scales. This paper aims to identify some of these key challenges and to highlight relevant research fields that might help to address them.

Industrial drying is one of the most energy inten-sive manufacturing processes and it is utilized across various industries, making it an ideal target for optimization. To achieve this condition based drying can be implemented, which requires knowledge of the product's internal moisture content (MC). This knowledge can be approximated through MC estimation. In this work, we frame MC estimation as a Time Series Extrinsic Regression (TSER) problem and investigate the performance of state-of-the-art TSER models. Additionally, insufficient training data is a major challenge in machine learning, especially for industrial applications, due to the prohibitively high cost of production line experiments. This acts as a barrier to the adoption of machine learning methods in industrial settings. In this work we propose a data augmentation method for TSER problems called Time Slicing. The proposed data augmentation method and TSER models are applied to a TSER dataset of industrial drying of bulky filter media products. It is shown that the TSER approach using an LSTM is able to out-perform the tabular data approach. Furthermore, it is shown that the proposed data augmentation method, when applied to the LSTM, improves its performance by 17.4% and 30.1% with regard to mean absolute error and mean squared error, respectively.