Abstract
With the growing adoption of machine learning systems in high-stakes domains such as healthcare, finance, and public administration, ensuring that these systems behave responsibly has become an urgent concern. Initiatives such as the EU AI Act and the broader movement toward responsible AI have highlighted fairness as a key challenge in the development and deployment of such technologies. Although existing tools support model op timization through hyperparameter tuning and algorithm selection, they often neglect the broader pipeline, overlooking how factors like data bias and model evaluation practices contribute to fairness. This paper presents Fairforge, a tool designed to support responsible ML development by guiding users through essential stages of the pipeline. These include data preprocessing with bias-awareness, fairness-informed model training, and postprocessing correction techniques. Fairforge also provides integrated interfaces for evalu ating both performance and fairness metrics. The tool aims to make state-of-the-art fairness techniques accessible to users without deep expertise in the field. To validate the effectiveness of Fairforge, we conducted a series of usability tests involving users with di verse levels of technical backgrounds. The results demonstrate that the tool helps promote fairness in model development, even among non-expert practitioners.

Abstract
Cardiovascular diseases remain the leading cause of morbidity and mortality worldwide. Continuous electrocardiographic (ECG) monitoring is essential for prevention and treatment, but conventional approaches based on the need for some voluntary action often limit comfort and adherence in long-term use. This study investigates the feasibility of acquiring ECG signals from a toilet seat interface embedding dry electrodes in the posterior thighs. A total of 30 hospitalised patients with diverse cardiovascular conditions-including arrhythmias, ischemic heart disease, heart failure, structural abnormalities, and aneurysms-were enrolled. Thigh-acquired ECGs were recorded simultaneously with conventional limb-lead signals and analysed for morphology, heart rate variability (HRV), and disease-related clustering. Thigh-based ECGs demonstrated clear P-QRS-T complexes with preserved morphology, allowing reliable extraction of mean templates and HRV metrics. The comparison between pathological and normal groups showed that post-surgical aortic repair patients had ECG profiles closest to the normal cluster; in contrast, aortic stenosis (AS) appeared most distant. HRV analysis revealed disease-specific autonomic patterns: patients with tricuspid or mitral involvement exhibited higher variability (SDNN up to 140 ms), whereas those with aortic valve disease presented markedly reduced parasympathetic indices (RMSSD and pNN50). Principal component analysis of multi-feature ECG data identified overlapping groups of Acute Coronary Syndrome, Unstable Angina and Ascending Aortic Aneurysm. At the same time, hierarchical clustering confirmed the distinct separation of conditions with severe hemodynamic disruption, such as PS and AS. These findings support the feasibility of unobtrusive thigh-based ECG monitoring via a toilet-seat interface, enabling reliable signal acquisition, HRV analysis, and preliminary patient stratification. This approach may lay the groundwork for future home-based cardiovascular screening and telemedicine applications.

Abstract
The pursuit of bio-inspired locomotion requires control architectures that are computationally efficient, deterministically synchronized, and modular. While centralized controllers offer simplicity, they suffer from wiring complexity. Consequently, distributed systems reduce cabling but introduce challenges in time synchronization due to communication latency. This paper presents an evaluation of a distributed control system for a quadrupedal robot utilizing the Raspberry Pi Pico 2W. An analysis was performed on a decentralized topology in which independent microcontrollers control specific parts of the robot rather than the entire system or a limb, and are synchronized via Micro-ROS over a serial transport layer. Leveraging the RP2350's dual-core symmetric multiprocessing (SMP) capabilities and FreeRTOS, the system maintains a 1 kHz motor control loop on Core 1, employing a 'Hybrid SpinWait' strategy to minimize scheduling jitter. To systematically test the architecture, the system underwent five stress tests: Core Isolation Jitter Analysis, Dual-Board Step Response Latency, Safety Disconnect validation, and Time-based execution. Experimental data reveal a transport-induced bottleneck capping telemetry at $\approx$ 400 Hz, despite internal generation >500 Hz, confirming that the USB-Serial interface is the primary limiter for dynamic gait synchronization, requiring a future transition to UDP-based Ethernet.

Abstract
Missing data is a common problem in real-world datasets and can be characterized as the lack of information on one or multiple variables in a dataset. The most frequent technique for handling this issue is imputation, which consists in the replacement of the missing values according to a predefined criterion. Since missing values are often imputed based on the known values in the dataset, existing data issues can be propagated during the imputation process. One such issue is fairness, a concept integral to responsible Artificial Intelligence practices. This work investigates the impact of the imputation process on system fairness by examining how imputation affects the fairness of predictions in Machine Learning models. It provides a comprehensive analysis covering thirteen unfair benchmark datasets with six state-of-the-art imputation strategies under synthetic Missing Not At Random and Missing At Random mechanisms in a multivariate scenario with 10%, 20%, 40%, and 60% of missing rates. Fairness was measured by the following metrics: Statistical Parity, Equalized Odds, Equality of Opportunity, Predictive Equality, Equality of Positive, and Negative Predicted Values. The results demonstrate that the missing mechanism, the classifier choice, and the imputation strategy decisively influence the fairness of the predictions obtained by the Machine Learning models.

Abstract
This special issue showcases state-of-the-art research at the intersection of analytics and retail operations. As the retail landscape becomes increasingly complex - driven by omnichannel strategies, evolving customer expectations, and a surge in data availability - analytics has emerged as a critical enabler of operational efficiency, customer experience, responsiveness, and sustainability and ethics. Collectively, these contributions demonstrate how advanced analytics can support retailers in navigating uncertainty, personalizing services, and scaling up innovation across formats and channels. The articles featured in this issue address a diverse set of decision domains, including warehousing, inventory and assortment planning, and distribution and last-mile delivery. Methodologically, they span descriptive, prescriptive, and hybrid approaches, leveraging tools such as machine learning, stochastic modeling, and dynamic optimization. By grounding models in real-world data and focusing on practical implementation, the issue provides actionable insights for both scholars and practitioners. It also highlights emerging opportunities for future research on behavioral integration, human-machine collaboration, and the ethical dimensions of retail analytics.

Abstract
In this paper, we introduce a novel approach for rigorously verifying safety properties of state machine specifications. Our method leverages an auto-active verifier and centers around the use of action functions annotated with contracts. These contracts facilitate inductive invariant checking, ensuring correctness during system execution. Our approach is further supported by the Why3-do library, which extends the Why3 tool's capabilities to verify concurrent and distributed algorithms using state machines. Two distinctive features of Why3-do are: (i) it supports specification refinement through refinement mappings, enabling hierarchical reasoning about distributed algorithms; and (ii) it can be easily extended to make verifying specific classes of systems more convenient. In particular, the library contains models allowing for message-passing algorithms to be described with programmed handlers, assuming different network semantics. A gallery of examples, all verified with Why3 using SMT solvers as proof tools, is also described in the paper. It contains several auto-actively verified concurrent and distributed algorithms, including the Paxos consensus algorithm.