Abstract

Abstract
Deep learning has been extensively applied to medical imaging tasks over the past years, achieving outstanding results. However, the obscure reasoning of the models and the lack of supportive evidence causes both clinicians and patients to distrust the models' predictions, hindering their adoption in clinical practice. In recent years, the research community has focused on developing explanations capable of revealing a model's reasoning. Among various types of explanations, example-based explanations emerged as particularly intuitive for medical practitioners. Despite the intuitiveness and wide development of example-based explanations, no work provides a comprehensive review of existing example-based explainability works in the medical image domain. In this work, we review works that provide example-based explanations for medical imaging tasks, reflecting on their strengths and limitations. We identify the absence of objective evaluation metrics, the lack of clinical validation and privacy concerns as the main issues that hinder the deployment of example-based explanations in clinical practice. Finally, we reflect on future directions contributing towards the deployment of example-based explainability in clinical practice.

Abstract
Aesthetic outcomes in plastic and oncological surgery play a fundamental role in restoring patients' self-esteem, social engagement, and overall quality of life. Yet, managing pre-operative expectations and objectively assessing post-operative results remain as difficult challenges, compounded by the subjective nature of beauty and the scarcity of standardized evaluation tools. To address these challenges, we conduct a systematic review assessing computational methods for the prediction and evaluation of the aesthetic outcomes of plastic and oncological surgery, adhering to PRISMA guidelines. We propose a goal-oriented taxonomy that partitions computational approaches into two main categories: (1) prediction methods that pre-operatively predict the results of surgery through retrieval-based systems, generative artificial intelligence and advanced 3D modeling techniques, and (2) evaluation strategies that assess the post-operative outcomes through objective measurements, traditional machine learning, and deep learning models. Our synthesis indicates a potential paradigm shift from early work that relied on manual image annotation and manipulation to recent research that predominantly employs artificial intelligence. Nevertheless, over 90% of datasets remain private, and validation processes diverge among techniques with similar goals, limiting reproducibility and fair methodological comparisons. We conclude by advocating for the creation of larger publicly accessible datasets, integration of vision-language models to capture patient-reported outcomes, and rigorous clinical validation to ensure equitable, patient-centered care. By bridging computational innovation with clinical practice, this study contributes towards a more transparent, reliable, and personalized aesthetic outcome prediction and assessment.

Abstract
The shift to AI-native 6G networks demands autonomous slicing strategies that can adapt to diverse and evolving edge and IoT service needs. Two paradigms have emerged: Learn to Slice (L2S), where AI optimizes network slicing for general services, and Slice to Learn (S2L), where slices support AI model training, often offloaded from Internet of Things (IoT) devices. Existing S2L approaches typically optimize communication or computation in isolation. This paper presents the first unified framework that jointly optimizes communication resources, computation capacity, and AI hyperparameters to maximize the average accuracy of multiple concurrent AI services. We address the complexity of this joint problem by applying L2S-inspired techniques to enhance S2L, introducing two autonomous agents: EXP3 from online convex optimization and DQN from deep reinforcement learning. Extensive experiments demonstrate and contrast the effectiveness of these agents in maximizing aggregated AI accuracy, supporting knowledge transfer, and sustaining robust performance under adversarial and long-term conditions, thereby enhancing the realization of zero-touch network management for AI services in 6G networks, supporting resource-constrained IoT.

Abstract
Background and objectives: Timely treatment of pediatric obstructive sleep apnea (OSA) can prevent or reverse neurocognitive and cardiovascular morbidities. However, whether distinct phenotypes exist and account for divergent treatment effectiveness remains unknown. In this study, our goal is threefold: i) to define new data-driven pediatric OSA phenotypes, ii) to evaluate possible treatment effectiveness differences among them, and iii) to assess phenotypic information in predicting OSA resolution. Methods: We involved 22 sociodemographic, anthropometric, and clinical data from 464 children (5-10 years old) from the Childhood Adenotonsillectomy Trial (CHAT) database. Baseline information was used to automatically define pediatric OSA phenotypes using a new unsupervised subject-based association network. Follow-up data (7 months later) were used to evaluate the effects of the therapeutic intervention in terms of changes in the obstructive apnea-hypopnea index (OAHI) and the resolution of OSA (OAHI < 1 event per hour). An explainable artificial intelligence (XAI) approach was also developed to assess phenotypic information as OSA resolution predictor at baseline. Results: Our approach identified three OSA phenotypes (PHOSA1-PHOSA3), with PHOSA2 showing significantly lower odds of OSA recovery than PHOSA1 and PHOSA3 when treatment information was not considered (odds ratios, OR: 1.64 and 1.66, 95 % confidence intervals, CI: 1.03-2.62 and 1.01-2.69, respectively). The odds of OSA recovery were also significantly lower in PHOSA2 than in PHOSA3 when adenotonsillectomy was adopted as treatment (OR: 2.60, 95 % CI: 1.26-5.39). Our XAI approach identified 79.4 % (CI: 69.9-88.0 %) of children reaching OSA resolution after adenotonsillectomy, with a positive predictive value of 77.8 % (CI: 70.3 %-86.0 %). Conclusions: Our new subject-based association network successfully identified three clinically useful pediatric OSA phenotypes with different odds of therapeutic intervention effectiveness. Specifically, we found that children of any sex, >6 years old, overweight or obese, and with enlarged neck and waist circumference (PHOSA2) have less odds of recovering from OSA. Similarly, younger female children with no enlarged neck (PHOSA3) have higher odds of benefiting from adenotonsillectomy.

Abstract
Context. Direct observations of exoplanet and brown dwarf companions with near-infrared interferometry, first enabled by the dualfield mode of VLTI/GRAVITY, provide unique measurements of the objects' orbital motions and atmospheric compositions. Aims. Here we compile a homogeneous library of all exoplanet and brown dwarf K-band spectra observed by GRAVITY thus far. This ExoGRAVITY Spectral Library is made publicly available online. Methods. We re-reduced all the available GRAVITY dual-field high-contrast data in a uniform and highly automated way and, where companions were detected, extracted their similar to 2.0-2.4 mu m K-band contrast spectra. We then derived stellar model atmospheres for all the employed flux references (either the host star or the swap calibrator), which we used to convert the companion contrast into companion flux spectra. Solely from the resulting GRAVITY K-band flux spectra, we extracted spectral types, spectral indices, and bulk physical properties for all the companions. Finally, and with the help of age constraints from the literature, we also derived isochronal masses for most of the companions using evolutionary models. Results. The resulting library contains R similar to 500 GRAVITY K-band spectra of 39 substellar companions from late M to late T spectral types, including the entire L-T transition. Throughout this transition, a shift from CO-dominated late M- and L-type dwarfs to CH4-dominated T-type dwarfs can be observed in the K-band. The GRAVITY spectra alone constrain the objects' bolometric luminosity to typically within +/- 0.15 dex. The derived isochronal masses agree with dynamical masses from the literature where available, except for HD 4113 c for which we confirm its previously reported potential underluminosity. Conclusions. Medium-resolution spectroscopy of substellar companions with GRAVITY provides insight into the carbon chemistry and the cloudiness of these objects' atmospheres. It also constrains these objects' bolometric luminosities, which can yield measurements of their formation entropy if combined with dynamical masses, for instance from Gaia and GRAVITY astrometry.