Abstract
The use of process improvement methodologies to assist and support the improvement of processes has proven to be an important mechanism for effectively implementing these improvements. However, there is difficulty in choosing the best methodology and to ensure that it will lead to the best improvement results. In this sense, the research questions of this work can be formulated as the following: H1-There are differences between the major process improvement methodologies and gaps not covered by them; H2-A new process improving methodology may mitigate the gaps identified in the existing process improvement methodologies. Comparing the main process improvement methodologies available in the literature, namely, PDCA, Six Sigma, DMAIC, QC Story, 8D, TOC and Lean, it was proven the research question H1. To validate the research question H2 a new process improvement methodology, the MAPIC, was then proposed and compared with the other methodologies. From a theoretical view point, the research question H2 was validated, because the MAPIC covers the existed gaps from the others methodologies, namely, that there is no phase to promote proactive continuous improvement, nor to validate the proposed improvement before its implementation. As for its practical validation, the MAPIC is being applied in a case study and the results will be presented in further work.

Abstract
Humour poses a unique challenge for artificial intelligence, as it often relies on non-literal language, cultural references, and linguistic creativity. The JOKER Lab, now in its fourth year, aims to advance computational humour research through shared tasks on curated, multilingual datasets, with applications in education, computer-mediated communication and translation, and conversational AI. This paper provides an overview of the JOKER Lab held at CLEF 2025, detailing the setup and results of its three main tasks: (1) humour-aware information retrieval, which involves searching a document collection for humorous texts relevant to user queries in either English or Portuguese; (2) pun translation, focussed on humour-preserving translation of paronomastic jokes from English into French; and (3) onomastic wordplay translation, a task addressing the translation of name-based wordplay from English into French. The 2025 edition builds upon previous iterations by expanding datasets and emphasising nuanced, manual evaluation methods. The Task 1 results show a marked improvement this year, apparently due to participants' judicious combination of retrieval and filtering techniques. Tasks 2 and 3 remain challenging, not only in terms of system performance but also in terms of defining meaningful and reliable evaluation metrics.

Abstract
Renewable Energy Communities (RECs) need performance-based methods to share locally generated energy to prevent free-riding, incentivize consumer behavior, and improve overall social well-being through sector interaction. We tackle the challenge of ranking REC members for local energy allocation factor purposes, based on multidimensional household waste sorting performance, where efficiency changes over time and trade-offs exist among waste streams. We created a ranking system that balances stability (for fairness) with responsiveness (to reward improvement), compensating the REC manager promoter (municipality). The method combines historical frontier analysis with Mahalanobis distance, following: (1) DEA-derived weights to combine inputs, (2) temporal frontiers for each waste stream, (3) projects current performance onto past benchmarks with a customized rolling window, (4) calculates multivariate z-scores through Mahalanobis distance, and (5) ranks members by their statistical distance from historical norms. The proposed methodology enhancement is verified with synthetic data from 30 households over 14 months, with 8 evaluation periods. It shows 71.4% rank category stability compared to 49.0% for monthly DEA, a 22.4 percentage point increase, while still detecting performance changes. The system accounts for output correlations, with mostly positive links between waste streams ((Formula presented) glass-packages, (Formula presented) glass-organic). Mahalanobis distance fairly rewards balanced performance across related dimensions. Sensitivity tests indicate that the approach is robust to variations in parameter choices. The framework provides a straightforward computational method (<1 s per evaluation) that yields rankings with statistical significance for consumer communication. It is the first framework designed specifically for temporal performance ranking in incentive allocation. © 2026 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC license. http://creativecommons.org/licenses/by-nc/4.0/

Abstract
PurposeIn an increasingly competitive market, equipment availability is a strategic variable for the competitiveness and success of companies. The objective of the research in this article is to present contributions to reduce unplanned production stoppages and optimise the operational efficiency of an injection moulding machine. This will be achieved by developing a systematic strategy to integrate predictive and condition-based maintenance systems with maintenance management software.Design/methodology/approachThe model developed is based on the continuous monitoring of electrical signals and vibrations, with the processing of data collected in real time through a script developed in Python. This integrates the information into the maintenance management software, facilitating a quick and accurate response to component wear conditions. The methodology employed was action research, as it was a case study developed in a real context, with active participation in development and implementation, with the aim of continuous improvement.FindingsIn August, a substantial increase was observed in the primary indicators: The mean time between failures (MTBF) increased by 97.36%, the mean time to repair (MTTR) increased by 313.31%, and the downtime was reduced by 65.04%. In December, although the figures were more moderate, significant improvements were maintained: The MTBF increased by 20%, the MTTR increased by 84%, and the downtime was reduced by 79%.Originality/valueThe findings of the study indicated that the implementation of a structured approach for the acquisition and monitoring of electrical signals and vibration data was imperative to achieve substantial gains.

Abstract
Machine learning (ML) frameworks, such as PyTorch and TensorFlow, rely on data loaders to preprocess data before feeding it to accelerators. When preprocessing is inefficiently pipelined, GPUs can remain idle over long periods of time, leading to substantial training delays. For example, PyTorch’s default data loaders can cause up to 76% GPU idleness. A key bottleneck is the variability in preprocessing time across samples within the same dataset. Existing data loaders are oblivious to this variability, training all samples uniformly. In this case, a single slow sample can stall the entire batch, causing head-of-line blocking. We present MinatoLoader, a general-purpose data loader for PyTorch that accelerates training and improves GPU utilization under single-server, multi-GPU settings. It continuously prepares data in background and constructs batches by prioritizing fast-to-process samples, while slower samples are processed in parallel. Experiments conducted over NVIDIA V100 and A100 GPUs show that MinatoLoader accelerates training by up to 7.5× (3.6× on average) over PyTorch DataLoader and Pecan, and up to 3× (2.2× on average) over DALI. It also increases average GPU utilization from 46% with PyTorch to 90%, while preserving model accuracy and enabling faster convergence. © 2026 Copyright held by the owner/author(s)

Abstract
Contrastive Analysis (CA) detects anomalies by contrasting patterns unique to a target group (e.g., unhealthy subjects) from those in a background group (e.g., healthy subjects). In the context of brain MRIs, existing CA approaches rely on supervised contrastive learning or variational autoencoders (VAEs) using both healthy and unhealthy data, but such reliance on target samples is challenging in clinical settings. Unsupervised Anomaly Detection (UAD) learns a reference representation of healthy anatomy, eliminating the need for target samples. Deviations from this reference distribution can indicate potential anomalies. In this context, diffusion models have been increasingly adopted in UAD due to their superior performance in image generation compared to VAEs. Nonetheless, precisely reconstructing the anatomy of the brain remains a challenge. In this work, we bridge CA and UAD by reformulating contrastive analysis principles for the unsupervised setting. We propose an unsupervised framework to improve the reconstruction quality by training a self-supervised contrastive encoder on healthy images to extract meaningful anatomical features. These features are used to condition a diffusion model to reconstruct the healthy appearance of a given image, enabling interpretable anomaly localization via pixel-wise comparison. We validate our approach through a proof-of-concept on a facial image dataset and further demonstrate its effectiveness on four brain MRI datasets, outperforming baseline methods in anomaly localization on the NOVA benchmark.