2 January 2026 in Open Research Europe
Background Maritime activity is expanding globally, increasing the demand for robust port security systems capable of detecting illegal trafficking. Due to the growing sophistication of smuggling methods, law enforcement agencies require advanced surveillance and prevention technologies such as those developed in the SMAUG project. In this context, initiatives such as the SMAUG project aim to deliver integrated surveillance capabilities coordinated by a high-level deep reinforcement learning (DRL) decision-making system that operates on image-based environmental representations. Despite their effectiveness, DRL models are closed-boxes, complicating continuous model monitoring (CMM). Conventional drift detection captures shifts in input or output distributions yet often fails to explain underlying problems. Explainable AI (XAI) techniques can provide a complementary approach with insights into the agent’s inner workings, enabling monitoring of the concept rather than just the data. Methods We propose FADMON, an XAI-driven concept drift detection method for image-based models. FADMON performs statistical drift tests on feature attributions to detect deviations in learned policies. We demonstrate how FADMON can enhance CMM with a three-stage model monitoring architecture that enables semi-supervised explainable model monitoring. We validate our approach with SMAUG’s decision-making DRL model on a simulated maritime port surveillance environment under multiple unforeseen scenarios. Results FADMON consistently flags drift on all drifted scenarios with mean p-values of 0.000 with no variance trough 30 repetitions, with lower mean p-values (0.553±0.215) on non-drifted scenarios with respect to other established drift detection methodologies such as prior probability shift detection (0.65 ± 0.000), though well above the standard 0.05 threshold. Conclusions FADMON can add an explainability layer to the monitoring system while also supporting detection of changes in the underlying interpretation of the input data by the model, monitoring the concept rather than the data, while matching established drift detection methods metrics-wise.
29 September 2025 in IEEE Signal Processing Letters
Deep Neural Network (DNN) based person re-identification (ReID) systems achieve high accuracy but operate as black boxes, limiting their deployment in high-risk scenarios such as forensic analysis and security screening. We propose an inherently interpretable and privacy-aware ReID framework built upon the SOLIDER-ReID architecture. For that, first, we apply B-cosification to the Swin Transformer backbone, replacing key layers with B-cos transforms to enable faithful, contribution maps that reveal the model's internal decision-making processes. Second, we introduce a multi-task learning module with attribute-specific heads that predict semantically meaningful soft biometrics (e.g., clothing type, colors, accessories), enhancing both performance and interpretability. Experiments on the Market-1501 Attribute dataset and the DukeMTMC-attribute-ReID dataset demonstrate that our approach achieves competitive performance while delivering near real-time, faithful explanations and maintaining privacy.
13 April 2026 in IEEE Signal Processing Letters
In-the-wild eye tracking datasets to train gaze vector estimation methods are noisy and unreliable, due both to limited image quality and to weak guarantees on label accuracy. At the same time, simply increasing dataset size does not guarantee better results, especially when diversity, balance and label reliability are not taken into account. In this work, we adopt a data-centric perspective and introduce an innovative data-pruning pipeline for robust gaze vector estimation. Our methodology leverages vision-language model embeddings and their zero-shot classification capabilities, together with an early-learning prediction-error analysis, to assign task-aware quality scores to individual samples. Moreover, we incorporate a region-weighted loss conditioned on gaze direction to better handle the uneven difficulty across the gaze space and to improve learning in under-represented regions. Finally, we validate our approach on a large relabeled subset of the GazeCapture dataset by training a hybrid gaze-estimation model, and show that a curated subset of 84K samples yields an accuracy improvement over a model trained on the full 2.45M-sample dataset (https://dx.doi.org/10.21227/c9ap-zm59 doi:10.21227/c9ap-zm59).