Expert Systems with Applications

The -Bonacci weighted sequence space {ell}(,) is introduced for [&ge;] 2, 1 [&le;] < {infty}, and > 0). It is proved to be a Banach space which is isometrically isomorphic to {ell}, with an unconditional Schauder basis. The space is also reflexive and uniformly convex. Its strict inclusion chain, continuous dual, and asymptotic truncation estimate of order - / are also established. Based on this framework, Scaled &#119898-BonacciNet is proposed as a lightweight CNN that rescales Fibonacci-Net channels using the dominant &#119898-Bonacci root and employs depth wise separable convolutions. Tribonacci-Net achieved 96.09% accuracy and a 0.9581 macro F1-score on the 15-class Brain MRI dataset using 0.85 million parameters and 0.366 GFLOPs. Pentanacci-Net attained 95.87% accuracy on the 44-class dataset while reducing parameters and FLOPs by 47.10% and 44.48%, respectively. On the CKD ultrasound dataset, Hexanacci-Net achieved 83.75% accuracy and a 0.7941 macro F1-score, with corresponding reductions of 50.37% and 45.10%. These findings support dominant-root scaling as a structured approach to lightweight CNN design.

In hypertension screening, previously diagnosed cases provide reliable positive labels, whereas individuals without a prior diagnosis comprise both undiagnosed hypertensive and non-hypertensive individuals, creating a canonical positive-unlabeled (PU) learning setting. We propose PU-Boost, a two-stage reliable-sample reconstruction framework that uses only readily obtainable demographic, behavioral, medical-history, and anthropometric variables. During training, only previously diagnosed hypertension cases are treated as labeled positives (P), while all remaining individuals are treated as unlabeled (U); no additional reference labels are supplied to the two-stage reconstruction procedure. Elastic Net first provides a global ranking of P versus U, after which a random forest refines the uncertain region and highly ambiguous observations are rejected. XGBoost is then trained on the reconstructed class-balanced pseudo-labeled sample. In the independent test set, PU-Boost achieved a sensitivity of 0.817 and a balanced accuracy of 0.682, the highest observed values among the compared models. It produced 729 false negatives, 41.1% fewer than standard XGBoost, although specificity decreased to 0.547. SHAP analysis indicated that established hypertension-related features, including age, body mass index, waist circumference, and family history, contributed strongly to model output. PU-Boost therefore provides a statistically motivated framework for reducing missed cases when screening for undiagnosed hypertension under incomplete diagnostic labeling.

Multiview data analysis has become increasingly important for integrating heterogeneous information from multiple data sources. Cooperative learning provides a flexible framework for multiview integration by bridging early and late fusion strategies. However, its use of squared loss makes it sensitive to outliers. In this study, we propose robust cooperative learning based on the soft-IPOD framework. The proposed method introduces an observation-specific outlier absorption vector into the cooperative learning objective, thereby reducing the influence of outlying observations while preserving information sharing across views. The resulting optimization problem can be formulated as a weighted Lasso-type problem using an augmented design matrix and can be naturally extended to more than two data views. Numerical simulations and a real-data application to labor onset prediction demonstrate the predictive performance of the proposed method in the presence of outliers.

Predicting implantation outcomes in vitro fertilization (IVF) is challenging because clinically relevant information becomes available at different points during treatment. In this study, we developed a stage-aware, leakage-conscious, and explainable machine-learning framework to predict implantation using a public IVF dataset. {beta}-HCG was excluded to avoid potential downstream target leakage. The data were divided into training, validation, and independent test sets. Several machine-learning models, including Logistic Regression, Random Forest (RF), Extra Trees, XGBoost, and Hist Gradient Boosting, were optimized using stratified cross-validation. A fixed RF model was then evaluated at three stages of the IVF process, using baseline couple characteristics, gamete and fertilization variables, and embryological variables, respectively. Model performance improved as additional treatment-related information became available, with the largest improvement observed after the inclusion of embryological variables. The final RF model showed good predictive performance on the independent test set. The number of embryos was consistently identified as the most important predictor across RF, permutation, and SHAP analyses. Overall, the findings suggest that IVF implantation prediction becomes more accurate as later-stage treatment information is incorporated, particularly information related to embryological characteristics. The proposed approach also provides interpretable predictions while reducing the risk of information leakage. Further evaluation using independent clinical datasets is needed before the framework can be considered for clinical use.

In many natural dynamical systems, tipping points emerge where slowly changing external factors induce shifts from one state to another, known as critical transitions or bifurcations. In epidemic systems, detecting these transitions is important for outbreak preparedness. Various statistical indicators like lag-1 autocorrelation and variance have been employed as early warning signals (EWS) of critical transitions by showing how the dynamics of the system exhibit critical slowing down (CSD) near the transition. More recently, deep learning-based algorithms have been developed that attempt to more accurately detect and classify bifurcations; however, they are limited by the systems simulated in the training dataset, the length of the time series, and have not been rigorously evaluated on real-world epidemic data spanning multiple geographic regions. In this paper, we evaluate existing deep learning-based bifurcation prediction models, on a CDC-curated dataset containing weekly influenza hospital admissions across all fifty U.S. states and the nation as a whole. We develop a pipeline for data preparation, reproduction number (Rt) estimation using the EpyEstim package (a Python implementation of the EpiEstim framework [4]), and bifurcation point identification. We systematically evaluate the models under several windowing strategies, including rolling, expanding, and fixed-length windows, and report accuracy, precision, recall, F1 score, and specificity. Our best-performing configuration (expanding window, length 100) achieves an accuracy of 90.09%, precision of 72.86%, recall of 96.23%, and F1 score of 82.93%. We additionally discuss Markov Regime Switching models as an alternative framework for detecting epidemic transitions. Our findings indicate that these models can generalize to real-world, state-level influenza data and are capable of online/real-time prediction.

Neuromarketing mainly seeks to enhance the prediction of marketing stimuli success, such as commercials and movie trailers, by integrating neurophysiological measures with traditional behavioral measures. In the current study, the authors tested whether neural activity could predict consumer valuation and how preferences change after watching commercials, both immediately and over time. Participants (n=161) watched images of products followed by commercials advertising those products while their neural activity was recorded using electroencephalograph (EEG). They watched the product images again a week later without EEG recordings. Immediately after each stimulus exposure, participants stated their willingness to pay (WTP) for the product and how much they liked the commercial. Standard EEG measures showed weak and inconsistent relationships with behavior and yielded near-chance predictions. In contrast, deep learning models applied to the raw neural data achieved substantially higher predictive accuracy, successfully predicting both immediate and delayed WTP and ad liking. Importantly, the authors were able to successfully predict preferences for new participants and/or new products the models were not trained on. However, accuracy declined as generalization demands increased. Moreover, prediction performance declined as value differences narrow. Together, these findings show that neural responses carry reliable and temporally persistent information about consumer valuation and its evolution over time.

Although largely preventable, cardiovascular disease (CVD) is the leading cause of death worldwide. Synthetic data generation (SDG) of health data can facilitate data sharing for research and development of personalised cardiovascular disease prevention tools. Current state-of-the-art SDG methods are limited to cross-sectional data or focused on short-term or in-patient EHR applications lacking benchmarks for generalisability. Since prevention is a long-term longitudinal objective, these limitations must be addressed. We evaluated two tabular generative adversarial networks, TGAN and CTGAN, for privacy and fidelity on a structured 6-year longitudinal trial dataset within CVD prevention. We compared performance with a rudimentary perturbation method. In addition, we performed a variable-wise evaluation and an age-stratified analysis of the higher performing model to assess preservation of clinically relevant characteristics. Overall, CTGAN showed better fidelity than TGAN. The models performed similarly within privacy, showing better results than the perturbation method. CTGAN had lower performance in comparison to the perturbation method within numerical fidelity of continuous variables and overall temporal fidelity. Furthermore, CTGAN's performance was lower for younger age groups than older. CTGAN shows better than or equal performance within fidelity and privacy than TGAN and the perturbation method on a long-term longitudinal dataset within CVD prevention. To the best of our knowledge, this is the first study evaluating CTGAN on a longitudinal dataset and our findings suggest that handling of temporal characteristics and imbalanced subgroups need further refinement.

Deep learning models for electroencephalography (EEG) pathology detection achieve competitive accuracy but offer limited interpretability. In this paper, we introduce concept bottleneck models (CBMs) for EEG pathology detection which use clinical EEG biomarkers computed analytically from the signal as a bottleneck to be used for prediction. Those concepts include band power, spectral ratios, alpha peak frequency and amplitude, hemispheric asymmetry and phase locking values. With the backbone based on shallow convolutional neural network (CNN), we compare a black box baseline with two variants of CBMs. The hybrid variant is a classifier receives a residual projection of the backbone features. The pure variant is a CBM where the classifier receives only the predicted concepts. Models are evaluated on abnormality detection and dementia classification followed by concept prediction quality and leakage analysis. Both CBMs approach the baseline on TUAB (82.7-82.9% vs 83.6%) and are 2-5 percentage points lower on CAUEEG (57.9-60.5% vs 62.7%). The pure CBM exhibits concept fidelity concern where the predicted concepts deviate from their ground truth values. Replacing them with their true values causes severe cost on accuracy by 26 and 22 points on TUAB and CAUEEG. The concept fidelity concern is mitigated by increasing the weight of the concept loss to match the ground truth concepts. the mitigation comes with small accuracy cost on TUAB but a relatively larger cost on CAUEEG. The code of the experiments and full logs are available at https://www.github.com/mhmdrdwn/cbm.

Background and Objective: Prescribing multiple drugs simultaneously, known as polypharmacy, is increasingly common in managing chronic diseases such as diabetes and cardiovascular conditions. Studies show that roughly 40% of elderly patients are on five or more medications, putting them at significant risk of adverse drug-drug interactions (DDIs) [1]. While computational methods based on Graph Neural Networks (GNNs) have shown strong potential for predicting these interactions from a drugs molecular structure, most existing models are limited in two important ways: they rely on attention mechanisms that cannot adapt to the unique context of each atom in a molecule, and they only predict whether an interaction exists rather than how dangerous it might be. Methods: We propose HMVA-DDI, a two-stage deep learning framework that addresses both shortcomings. In the first stage, our model predicts whether two drugs will interact. If it detects an interaction, the second stage classifies it into one of three clinical severity levels: minor, moderate, or major. To better capture molecular structure, we replace the standard attention mechanism with a more powerful version called Graph Attention Network v2 (GATv2), which considers each atoms specific context when learning its representation. We also introduce a technique to prevent information loss in deeper network layers, and we tailor the input features separately for each prediction task. Results: Evaluated on two standard benchmarks, DrugBank 5.0.3 [2] and DDInter [3], HMVA-DDI achieves an AUROC of 0.9819 and AUPRC of 0.9759 on binary interaction detection, a marked improvement over the baseline. For severity classification, our best model variant (qcSvr4) reaches an overall AUROC of 0.9444, with a particularly large improvement in detecting low-severity (Minor) interactions, a historically difficult class. Conclusion: HMVA-DDI bridges the gap between computational drug research and clinical usefulness by not only predicting interactions accurately but also indicating interaction severity. This two-level system can help healthcare professionals prioritize warnings, reducing the alert fatigue problem where doctors are overwhelmed by too many non-critical notifications.

Metabolic dysfunction-associated steatotic liver disease (MASLD) is highly prevalent among individuals with type 2 diabetes (T2D), yet accurate identification from electronic health records (EHRs) remains challenging because clinically relevant information is fragmented across structured data and unstructured clinical documentation. We developed a rule-based computable phenotype (CP) framework that integrates structured EHRs with clinical notes for scalable MASLD identification in patients with T2D. The framework comprises five complementary CP algorithms derived from validated fibrosis indices and liver injury biomarkers, together with a lightweight natural language processing module that identifies MASLD-related evidence using keyword matching, contextual filtering, and negation detection. The proposed framework was systematically evaluated across multiple threshold settings, with additional validation through age- and sex-stratified analyses and manual chart review. The findings demonstrate that integrating structured and unstructured EHR data provides an accurate and scalable framework for real-world MASLD phenotyping and offers a practical strategy for developing CPs for other chronic diseases.

In this work, we investigate the homogeneity of performance in binary classification between the Rotterdam phenotypes and phenotype prediction. For this purpose, a multi-task neural network architecture is designed to improve the prediction accuracy of phenotypes while maintaining the accuracy of binary classification. Using a clinical PCOS dataset which contains unlabeled samples of patients that cannot be attributed to any of the known phenotypes, three models are evaluated: optimized binary Random Forest (RF), six-class weighted class RF, and multi-task neural network (with a common hidden layer and two heads for binary and six-class classifications) optimized by a genetic algorithm. Techniques used for evaluation include ten-fold out-of-fold predictions, paired McNemar tests, bootstrap tests, ablation study, and clustering (Adjusted Rand Index, ARI). Binary RF model reached 90.2% accuracy; however, recall of different phenotypes was significantly unbalanced - from 75.0% of phenotype B to 97.8% of phenotype A. This difference of 22.8 % points was observed (23.3 points) even after excluding phenotype-specific features. While direct phenotype prediction improved recall for common phenotypes, it decreased recall for phenotype B to 43.8% and unclassified cases to 31.5%. Our multi-task model preserved 90.2% binary accuracy and significantly recovered phenotype B recall to 87.5% and unclassified cases to 79.6% (p<0.05). The cluster analysis confirmed the weak correlation between the results and labels (ARI=0.045). The ablation analysis demonstrates that feature selection and class weighting drive this improvement, isolating the specific components responsible for performance gains. We establish that aggregate binary accuracy does not reflect the heterogeneous detection of PCOS phenotypes. Direct phenotype prediction alone redistributes the classification errors, but the multi-task architecture successfully improves the detection of rare phenotypes without compromising binary accuracy. Phenotype-disaggregated evaluations, supported by significance testing and ablation, are essential for clinical classifier research.

Background: Accurate patient understanding of lung cancer screening (LCS) results is critical for engagement and follow-up adherence, given the currently low screening uptake. While artificial intelligence (AI) systems show promise in facilitating patient communication, its effective use in clinical settings requires appropriate invocation of imaging tools and self-regulation by deferring certain questions to physicians. To this end, we developed and evaluated a planner-executor style multimodal agentic system to answer simulated patient questions about LCS CT results. Methods: This retrospective study utilized 116 LCS CT reports and images collected from a tertiary academic hospital. The system employed a multi-agent architecture (planner and executor) to mimic clinical reasoning. We developed a large language model (LLM)-based question generation framework to prepare a comprehensive set of 699 simulated patient questions, categorized as answerable and defer-to-doctor to assess self-regulation. Performance was evaluated on tool-calling, self-regulation accuracy (ability to correctly defer out-of-scope questions to a human provider) assessed by LLM-as-a-judge, and clinical quality assessed by a reader performance study with four physicians evaluating a subset of 100 responses. Results: The system demonstrated high overall tool-calling accuracy of 93.1% (646/694; 95% CI: 91.2%, 95.0%) and self-regulation accuracy of 92.6% (462/499; 90.2%, 94.8%). For answerable questions, the percentage of responses receiving perfect 5.0 scores across all readers included 78.9% (180/228) for clinical accuracy (inter-rater agreement: Gwet's AC2=0.92), and 73.2% (167/228) for patient understandability (0.86). For defer-to-doctor questions, the percentage of responses included 77.3% (133/172) for clinical accuracy (0.80) and 70.3% (121/172) for patient understandability (0.78). Conclusions: The planner-executor style multi-modal agentic system reliably answers patient-specific questions about LCS. Its high tool-calling and self-regulation performance demonstrates its potential as a safe and effective digital communication facilitator in LCS programs.

Background and purpose: Accurate identification of cardiac MRI volume orientations is essential for reliable image interpretation and for enabling downstream automated analysis pipelines. However, orientation labels and commonly assigned manually, making the process time-consuming and prone to variability. Recent advances in vision-language models offers new opportunities for automated label generation by jointly modeling visual and semantic information. Methods: In this study, we systematically compare a conventional end-to-end transformer-based captioning model with a CLIPCap architecture for automated cardiac MRI view labelling. Both models were evaluated on a dataset of 1184 de-identified cardiac MR images spanning across six clinically relevant orientations: short axis, two-, three-, four-chamber views, left and right ventricular outflow tract views. The caption transformer was trained from scratch, whereas CLIPCap leveraged pretrained CLIP image embeddings combined with lightweight mapping network and a frozen language model. Results: Experimental results demonstrate a substantial performance advantage from CLIPCap, which achieved an overall accuracy of 98% compared to 68% for the caption transformer. CLIPCap consistently delivered high precision and recall across all orientation classes, including those of high clinical importance, while the end to end transformer exhibited unstable performance and complete failure in certain views. Conclusion: These findings highlight the benefits of leveraging large scale multimodal pretrained model for medical image labeling tasks, particularly in data limited settings. The study suggests that CLIPCap provides a more robust and clinically reliable solution for automated cardiac MRI orientation labeling, supporting its integration into efficient and scalable clinical workflows.

Dynamic systems often evolve faster than reliable measurements can be collected. More frequent indirect observations may fill these gaps, but their noise and differing scales complicate estimation of the underlying state. We developed Bayesian Anchored Latent Learning (BALL) to combine sparse reference measurements with denser records on an interpretable scale. A bidirectional transformer teacher reconstructs trajectories from complete training histories, and a forward-only transformer student learns from those reconstructions and observed measurements. Across 30 simulations, BALL reduced composite trajectory error by 7% relative to the same transformer trained directly from measurements, achieving lower error across all datasets. In an applied example using psychiatric records from 1,609 patients, BALL improved the prediction of withheld questionnaire scores compared with direct transformer training at 21- and 28-day measurement schedules. Requests guided by estimated uncertainty further improved prediction. BALL connects retrospective learning with prospective state estimation and provides a framework for directing additional measurement.

Accurate breast lesion segmentation in ultrasound is difficult because speckle noise, heterogeneous lesion appearance, and poorly defined margins can obscure clinically relevant boundaries. This work presents UGB-SegNet, an uncertainty-guided breast ultrasound segmentation network that combines four components: multi-scale convolutional block attention (CBAM) at three EfficientNet-B0 encoder stages, a learnable feature pyramid network for adaptive multi-scale fusion, an entropy-guided decoder that routes fine-grained encoder information toward uncertain regions, and a boundary-aware composite loss with deep supervision. The model was evaluated on the public BUSI dataset (647 annotated lesion images) using a stratified 454/64/129 train/validation/test split. On the held-out test set, UGB-SegNet achieved 81.30% Dice, 68.49% IoU, 81.71% sensitivity, 98.14% specificity, and 15.63-pixel HD95. Under identical training conditions it outperformed five directly trained baselines: U-Net, Attention U-Net, SegNet, DeepLabV3+, and TransUNet. A five-fold resampling experiment over the development set with the held-out test set fixed yielded a mean Dice of 76.36% +/- 1.66% (95% CI 74.90-77.81). Reported ablations attribute gains of +2.06 Dice points to multi-scale CBAM, +1.52 to learnable FPN fusion, +1.22 to the uncertainty-guided decoder, and +1.24 to boundary-aware training. UGB-SegNet contains 14.00M parameters and runs at 62.4 FPS at 224 x 224 resolution on the tested NVIDIA GPU, providing a compact accuracy-efficiency trade-off for future prospective clinical validation.

Background: Attention-Deficit/Hyperactivity Disorder (ADHD) is a common neurodevelopmental disorder characterized by persistent symptoms of inattention, hyperactivity, and impulsivity that significantly affect cognitive, academic, and social functioning. Although electroencephalography (EEG) has emerged as a promising non-invasive tool for identifying objective neurophysiological biomarkers of ADHD because of its high temporal resolution and cost-effectiveness, many existing machine learning studies rely on single-domain EEG features, which may not adequately capture the complex functional organization of the brain. Methods: This study proposes a subject-level EEG classification framework that integrates three complementary feature domains: spectral characteristics, weighted Phase Lag Index (wPLI)-based functional connectivity, and topological descriptors derived from persistent homology. EEG recordings from 121 subjects were preprocessed and segmented into overlapping windows for feature extraction. Window-level features were aggregated to generate subject-level representations. Recursive Feature Elimination (RFE) was employed to select the 110 most informative features from an initial set of 1,043 descriptors. Five machine learning classifiers, including Support Vector Machine (SVM), Random Forest, XGBoost, CatBoost, and Multi-Layer Perceptron (MLP), were evaluated using stratified five-fold cross-validation. An ablation study was further conducted to assess the contribution of each feature domain. Model interpretability was investigated using SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME), two widely adopted Explainable Artificial Intelligence techniques for interpreting machine learning models. Results: Among the evaluated classifiers, the SVM achieved the highest classification performance, yielding an accuracy of 80.17%, an F1-score of 80.95%, a balanced accuracy of 80.14%, a receiver operating characteristic area under the curve (ROC-AUC) of 86.72%, and a precision recall area under the curve (PR-AUC) of 87.34%. The ablation study demonstrated that combining spectral, functional connectivity, and topological features consistently outperformed individual feature domains, highlighting the complementary nature of these representations. Explainable AI analyses further identified functional connectivity features as the primary contributors to classification performance while confirming the additional value of spectral and topological descriptors. Conclusions: The proposed framework demonstrates that integrating spectral analysis, functional connectivity, and topological data analysis provides an effective and interpretable approach for subject-level ADHD classification from EEG recordings. The combination of recursive feature selection, multi-domain feature fusion, ablation analysis, and Explainable Artificial Intelligence establishes a transparent and reproducible pipeline that may facilitate the development of reliable EEG-based clinical decision-support systems for ADHD.

In recent years, the issue of reproducibility in scientific studies has been at the forefront of scientific discussion. Many of these issues arise due to poor employment of statistical methods, making it crucial to address these knowledge gaps for frequently used methods. Elastic net regression (ENET) is one such algorithm used frequently in biology and public health due to its ability to select important predictors (also known as "features") from large data sets. However, ENET struggles with consistent feature selection and hyperparameter choice in these applications due to low observation counts and noisy data. This limitation hinders ENET's effectiveness in biological research due to low model reproducibility and predictor interpretability. We propose the Instability Quotient Criterion (IQC) as a metric to quantify the consistency of selected features for feature selection models such as ENET, with a specific emphasis on biological applications. IQC leverages ensemble modeling and Shannon's entropy in order to determine the stability of selected features derived from ENET models. We demonstrate IQC's robustness via simulation studies, showing that IQC is capable of detecting significant decreases in model stability with lower observation counts, higher dimensionality, and higher proportions of noisy signal. We additionally utilize simulation to provide a naive threshold for low versus high stability. Finally, we applied IQC to a case study to illustrate how our method can be integrated into a larger analysis pipeline when looking at microarray gene expression data. IQC indicated stability issues for differentially-expressed genes when classifying acute leukemia subtypes in spite of high classification success. Lastly, we discuss how model instability impedes interpretation of biologically significant genes for discerning cancer subtypes. Overall, we present IQC as a robust evaluation metric that can easily be incorporated into high-dimensional biological analyses.

The Hierarchical Gaussian Filter (HGF) is a hierarchical Bayesian model of learning under uncertainty and volatility, with a widely used MATLAB implementation. We present HGFX 1.0.0, a Python/JAX toolbox whose primary objective is validated behavioral and numerical compatibility with a frozen HGF Toolbox 8.2.0 reference, while removing MATLAB as a user-runtime dependency. Reimplementation is treated as a validation problem rather than source translation: configuration semantics, parameter transforms, forward trajectories, observation likelihoods, objectives, fitting and statistical surfaces, simulation workflows, official demo behavior, model selection, and backend agreement are compared against the pinned oracle under predeclared tolerances. Two official MATLAB demo workflows reproduce the reference at frozen trajectory tolerances, including a regime in which classic HGF encounters negative posterior precision while eHGF completes successfully. In paired model-selection validation, all 36 BIC winners agree between MATLAB and HGFX. Physical NVIDIA GPU applicability was confirmed on two Tesla T4 devices, with a maximum CPU-versus-GPU final-objective difference of 1.42e-14 against a frozen 1e-7 criterion. Historical parameter recovery results that did not meet predeclared criteria, together with exact MATLAB/HGFX limitation matches, are preserved and are not reclassified as scientific success. A prospectively gated comparison with pyhgf 0.3.2 shows binary64-scale agreement on mapped perceptual trajectories in one authorized three-level binary-HGF cell, while participant-response negative log-likelihood is retained as not directly comparable. HGFX therefore provides a MATLAB-independent Python implementation with an explicit evidence model that separates direct parity, matched reference limitations, backend applicability, and future performance claims.

Traditional cell tracking methods focus on fully automated tracking. The quality of the generated tracks, however, is heavily influenced by factors such as cell speed and density. In complex scenarios where track correctness is essential, an efficient and reliable tool for manual track correction remains a key component of the cell tracking process. Most tracking tools that allow for such manual correction do so only as a post-processing step, using human feedback to correct one track error at a time -- without improving the quality of other tracks in the dataset. This highlights the need for a more efficient track correction method that iteratively integrates human feedback into the automated tracking process. In this paper we propose a human-in-the-loop approach to cell tracking and investigate the extent to which a tracking algorithm can learn from human input during the tracking process. We combine a graph representation of cell detections, Dijkstra's shortest path algorithm, and a neural network that predicts cell displacement to generate initial track suggestions. We introduce a tool that enables human feedback on individual tracks suggested by the algorithm; this feedback is then propagated to improve subsequent track generation. Compared to the traditional approach of fully automated tracking with correction as a post-processing step, our human-in-the-loop tracking framework reduces the number of corrections required. Finally, we show how this approach can be integrated with existing state-of-the-art tracking algorithms by using pre-generated tracks as a prior, improving the quality of tracks suggested to the human annotator.

Background. In people with type 2 diabetes (T2D), postprandial glucose weighs on glycaemic control, and carbohydrate counting is difficult. Patients can describe their meal but rarely quantify it. Objective. To assess (1) whether a small open large language model (LLM), fine-tuned on public data, estimates the carbohydrate content of a meal described in words, and (2) what this carbohydrate adds to the prediction of the postprandial glucose rise in two public cohorts. Methods. Step 1: the Qwen3-4B LLM was fine-tuned with QLoRA on Nutrition5k dishes and compared with an automatically matched USDA food composition table and with Qwen3-4B and Qwen3-8B used without training, on test dishes (internal control) and on 100 real meals of ShanghaiT2DM patients, whose reference carbohydrate was computed from the recorded quantities. Step 2: the maximum glucose rise between 0 and 120 minutes ({Delta}Gmax) was predicted by personalised gradient boosting (M2), compared with a spline linear regression and with the Chronos-2 foundation model, using only information available at meal start and person-level cross-validation, in ShanghaiT2DM (people with T2D recruited in two Shanghai hospitals; carbohydrate estimated by the LLM) and CGMacros (free-living adults in Santa Barbara, California; carbohydrate entered in the dataset, no text). Step 3: the model was transferred between datasets. 95% confidence intervals (CI) were obtained by cluster BCa bootstrap (2,000 resamples). Results. On the 100 real meals, the carbohydrate mean absolute error (MAE) of the fine-tuned model was 12.2 [9.9, 15.2] g, almost half that of the table (22.6 g; paired difference -10.5 [-16.8, -6.0] g) and nearly a third of that of the same model without fine-tuning (34.3 g), with a bias of -10.3 g; on test dishes it was 2.2 [2.0, 2.5] g. Personalised gradient boosting (M2) reached a correlation between predicted and observed rises of r = 0.63 [0.55, 0.69] in CGMacros (1,286 meals, 45 adults) and 0.49 [0.42, 0.54] in ShanghaiT2DM (2,339 meals, 99 people with T2D), about twice the linear reference (r = 0.32 and 0.23); for a rise above 50 mg/dL, its area under the ROC curve (AUROC) was 0.80 and 0.74, versus 0.64 and 0.68 with pre-meal glucose alone. Adding carbohydrate to pre-meal glucose changed r by 0.095 [0.068, 0.128] in CGMacros and by 0.004 [-0.010, 0.017] in ShanghaiT2DM; difference between datasets: 0.091 [0.060, 0.126]. Applied from one dataset to the other, the model lost correlation ({Delta}r = -0.11 [-0.17, -0.05] towards ShanghaiT2DM and -0.12 [-0.19, -0.04] towards CGMacros). The full "text [-&gt;] LLM [-&gt;] glucose" chain was evaluated in ShanghaiT2DM only. Conclusion. An open 4-billion-parameter LLM, fine-tuned with QLoRA on public data and usable offline, almost halves the error of carbohydrate counting for real meals described in words compared with an automatically matched table, with an underestimation of large meals. Personalised gradient boosting, based on pre-meal glucose and the person's usual response, doubles the correlation of a linear reference for predicting the glucose rise in both cohorts. Carbohydrate improved this prediction in CGMacros but not in ShanghaiT2DM; the two datasets differ in many respects, and the cause of this gap is not established.