JMJackie Ma

Jackie Ma

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Jackie Ma has a total of 142 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    From law to practice: model-level operationalization of technical governance requirements for high-risk AI in healthcare

    by Emílio Dolgener Cantú, David Mack, Katharina Weitz, Eva Weicken, Jackie Ma

    The European Union published the Artificial Intelligence Act (AI Act) in a strategic effort to align the development and application of Artificial Intelligence (AI) in the common market. This act, one of the first laws of its kind, strengthened the development of trustworthy and human-centered AI through new regulations, prescribing requirements that providers of AI systems must conform to. The Act prescribes model-level technical requirements for high-risk AI systems, such as in terms of accuracy, robustness and fairness, which are to be addressed and reported in the form of conformity assessments. We identify, however, a significant gap between the high-level legal descriptions provided in the law and the low-level requirements needed for the implementation of automated assessment tools for algorithmic auditing. This gap is a major obstacle in the operationalization of AI Act compliance. The difficulties are aggravated due the lack of harmonized AI standards and the existing heterogeneity of AI models, datasets, and tasks. To address this gap, we present an open-source proof-of-concept assessment tool to operationalize the technical requirements of the AI Act according to the current regulatory landscape. With this tool, we demonstrate a bottom-up interpretation of the technical requirements of the AI Act for an automated evaluation of accuracy, robustness, and fairness. We apply our work to two neural network-based healthcare applications and discuss the requirements for further integration to initial conformity assessment and to extended application to post-market monitoring and other algorithmic auditing frameworks. With further work, the base structure presented here can be applied for integration beyond healthcare to other AI high-risk application domains.

  • Open Access

    Structural Compactness as a Complementary Criterion for Explanation Quality

    by Mohammad Mahdi Mesgari, Jackie Ma, Wojciech Samek, Sebastian Lapuschkin, Leander Weber

  • Open Access

    Benchmarking Vector Quantized Auto-Encoders for Multimodal Vision-Language Tokenization of Medical Images

    by Emílio Dolgener Cantú, Manasi Acharya, Jim Berend, Md Golam Rasul, Jackie Ma

    Medical diagnosis often benefits from the side-by-side analysis of different sources of information, motivating also the development of multimodal approaches in AI-based applications for healthcare. A current technique for vision-language multimodality is a two-step process of (1) image tokenization through vector quantization followed by (2) auto-regressive modeling. In this work, we explore the image tokenization process for datasets in the medical domain, which are characterized by smaller scale and narrower semantic coverage than natural image datasets. We benchmark both reconstruction quality, codebook collapse and representational capacity of VQ-VAEs across a variety of settings, surpassing state of the art in the reconstruction task and providing a stepping stone for further development of medical multimodal auto-regressive techniques.

  • Open Access

    From fires to floods: A participatory case study on stakeholder requirements for AI-supported satellite services in disaster management

    by Katharina Weitz, Alen Berta, Noelia Otero, Monique M. Kuglitsch, Ivanka Pelivan, Jackie Ma, Pierre Philippe Mathieu, Stephane Pirio

  • On the Predictive Skill of Artificial Intelligence–Based Weather Models for Extreme Events Using Uncertainty Quantification

    by Rodrigo Almeida, Noelia Otero, Miguel-Ángel Fernández-Torres, Jackie Ma

    Accurate prediction of extreme weather events remains a major challenge for artificial intelligence–based weather prediction systems. While deterministic models such as FuXi, GraphCast, and spherical Fourier neural operator (SFNO) have achieved competitive forecast skill relative to numerical weather prediction, their ability to represent uncertainty and capture extremes is still limited. This study investigates how state-of-the-art deterministic artificial intelligence–based models respond to initial-condition perturbations and evaluates the resulting ensembles in forecasting extremes. Using four perturbation strategies (Gaussian, Perlin noise, hemispheric centered bred vectors, and huge ensembles), we generate 50-member ensembles for the August 2022 Pakistan floods and China heat wave and complement these case studies with a global threshold-based evaluation. Ensemble skill is assessed against ERA5 and compared with Integrated Forecasting System Ensemble (IFS ENS) and the Artificial Intelligence Forecasting System Ensemble (AIFS ENS) probabilistic models using deterministic and probabilistic metrics. Results show that simpler perturbations like Gaussian and Perlin noise produce similarly realistic ensemble spread and probabilistic skill as flow-based approaches like hemispheric centered bred vector (HCBV) and huge ensembles (HENS), narrowing but not closing the performance gap with numerical weather prediction ensembles or native probabilistic models which retain the highest probabilistic skill across variables. Model choice is the dominant factor for ensemble performance, not the perturbation method. Across variables, models capture temperature extremes more effectively than precipitation. These findings demonstrate that simple input perturbations can extend deterministic models toward probabilistic forecasting in hardware-constrained settings, supporting artificial intelligence–driven early warning systems. Significance Statement: Artificial intelligence–based weather models can rival traditional numerical models in accuracy and run faster, but most issue a single deterministic prediction without telling users how confident it is—a critical limitation for extreme events. We test whether perturbing the inputs of these models can turn them into ensembles that quantify uncertainty. We find that the choice of underlying model matters more than the choice of perturbation strategy and that perturbation-based ensembles do not match the skill of dedicated probabilistic systems. Input perturbations are a useful but limited tool, most relevant when hardware constraints preclude running dedicated probabilistic systems, and one of the several directions toward reliable artificial intelligence–driven early warning systems.


  • Open Access

    From Fires to Floods: Designing Human-Centered AI through Participatory Workshops in Disaster Management

    by Katharina Weitz, Alen Berta, Noelia Otero Felipe, Monique M. Kuglitsch, Ivanka Pelivan, Jackie Ma, Pierre Phillippe Mathieu, Stephane Pirio

    AI-based systems that use Earth observation (EO) and satellite communication (SatCom) services can support disaster risk management. However, successful deployment in the field requires ongoing stakeholder involvement throughout system development. Human-Centered AI (HCAI) provides a suitable framework for this by explicitly addressing human needs and requirements. In this work, we present an HCAI-inspired approach that connects stakeholder participation to the design of a demonstrator. We conducted participatory online workshops with stakeholders to understand their roles, goals, tasks, and requirements. These insights guided the development of an AI-supported prototype integrated into a multi-domain marketplace demonstrator that combines EO and SatCom services. The first prototype iteration was evaluated in two use cases: burned area detection and flood modeling, indicating differing access to EO and SatCom services across stakeholder groups but broadly similar goals and strong interest in integrated, multi-modal services, alongside concerns about real-time data and user training.

  • Open Access

    Spatially resolved meteorological and ancillary data in Central Europe for rainfall streamflow modeling

    by Marc Aurel Vischer, Noelia Otero, Jackie Ma

    We present a gridded dataset for rainfall streamflow modeling that is fully spatially resolved and covers five complete river basins in central Europe: upper Danube, Elbe, Oder, Rhine, and Weser (https://doi.org/10.4211/hs.d7f2cbb587ab4a75ac7987854e8f62ca, Vischer et al., 2025a). We compiled meteorological forcings and a variety of ancillary information on soil, rock, land cover, and orography. The data is harmonized to a regular 9 km × 9 km grid, temporal resolution is daily from 1980 to 2024. We also provide code to further combine our dataset with publicly available river discharge data for end-to-end rainfall streamflow modeling. We have used this data to demonstrate how neural network-driven hydrological modeling can be taken beyond lumped catchments, and want to facilitate direct comparisons between different model types.


  • Open Access

    The Clinician Model Card: development and evaluation of clinician-centered documentation for AI-based clinical decision support

    by Louis Agha-Mir-Salim, Nicolas Frey, Zoe Kaiser, Lina Mosch, Eva Weicken, Oscar Freyer, Jackie Ma, Mirja Mittermaier, Alexander Meyer, Stephen Gilbert, et al.

    AI documentation frameworks remain poorly designed for point-of-care use, leaving clinicians without actionable information on how to use clinical AI models when they need it most. We developed the Clinician Model Card, an interactive, clinician-centered documentation tool, and evaluated it in a sequential exploratory mixed-methods study: interviews with 12 physicians informed iterative co-design, evaluated in a national survey of 129 physicians across Germany. The tool was well-received: 84% agreed it should be routinely available, and 66% considered its content relevant to clinical decision-making. Yet comprehensibility of statistical performance metrics remained poor despite targeted interventions: only 32% understood the Validation & Performance section well, and fewer than 54% correctly interpreted AUROC or PPV, with AI literacy as strong predictor of comprehension (ρ = 0.59). We propose empirically derived design principles for clinician-centered AI documentation. Effective AI transparency requires not only clinician-friendly design and workflow integration, but sustained investment in AI literacy.


  • Open Access

    Multimodal Deep Learning for Prediction of Progression-Free Survival in Patients with Neuroendocrine Tumors Undergoing 177Lu-Based Peptide Receptor Radionuclide Therapy

    by Simon Baur, Tristan Ruhwedel, Ekin Böke, Zuzanna Kobus, Gergana Lishkova, Christoph Wetz, Holger Amthauer, Christoph Roderburg, Frank Tacke, Julian M. Rogasch, et al.

    Background/Objectives: Peptide receptor radionuclide therapy (PRRT) is an established treatment for metastatic neuroendocrine tumors (NETs), yet long-term disease control occurs only in a subset of patients. Predicting progression-free survival (PFS) could support individualized treatment planning. This study evaluates laboratory, imaging, and multimodal deep learning models for PFS prediction in PRRT-treated patients. Methods: In this retrospective, single-center study 116 patients with metastatic NETs undergoing [177Lu]Lu-DOTATOC were included. Clinical characteristics, laboratory values, and pretherapeutic somatostatin receptor positron emission tomography/computed tomographies (SR-PET/CTs) were collected. Seven models were trained to classify low- vs. high-PFS groups, including unimodal (laboratory, SR-PET, or CT) and multimodal fusion approaches. Performance was assessed via repeated 3-fold cross-validation with area under the receiver operating characteristic curve (AUROC) and area under the precision–recall curve (AUPRC). Explainability was evaluated by feature importance analysis and gradient based saliency maps. Results: Forty-two patients (36%) displayed short PFS (≤1 year) and 74 patients displayed long PFS (>1 year). Groups were similar in most characteristics, except for higher baseline chromogranin A (p = 0.003), elevated γ-GT (p = 0.002), and fewer PRRT cycles (p < 0.001) in short-PFS patients. The Random Forest model trained only on laboratory biomarkers reached an AUROC of 0.59 ± 0.02. Unimodal three-dimensional convolutional neural networks using SR-PET or CT performed worse (AUROC 0.42 ± 0.03 and 0.54 ± 0.01, respectively). A multimodal fusion model integrating laboratory values, SR-PET, and CT—augmented with a pretrained CT branch—achieved the best results (AUROC 0.72 ± 0.01, AUPRC 0.80 ± 0.01). Explainability analyses provided insights into model predictions, with explainability patterns in the fusion model appearing physiologically plausible and predominantly tumor-focused. Conclusions: Multimodal deep learning combining SR-PET, CT, and laboratory biomarkers outperformed unimodal approaches for PFS prediction after PRRT. Upon external validation, such models may support risk-adapted follow-up strategies.


  • Open Access

    Deep Learning-Based Precipitation Nowcasting for Operational and Flash-Flood Applications

    by Rodrigo Almeida, Jamil Göttlich, Noelia Otero, Marian Jurasek, Ladislav Méri, Zinaw Dingetu Shenga, Aitor Atencia, Jackie Ma

    Accurate short-term precipitation nowcasting is crucial for disaster risk reduction, flash-flood early warning, and water resource management. Conventional nowcasting approaches, such as extrapolation-based radar methods or numerical weather prediction models, often struggle to capture the nonlinear evolution of convective systems and are computationally demanding for rapid updates at high spatial and temporal resolution. The ability to provide reliable high-resolution forecasts at lead times of minutes to hours is particularly important for mitigating the societal and economic impacts of intense rainfall events. Recent developments in deep learning (DL), in combination with high-resolution radar observations, represent a compelling alternative for improving short-term precipitation forecasting. Radar-based precipitation data are particularly well suited for nowcasting applications due to their fine spatio-temporal resolution and ability to capture the dynamic structure and movement of precipitation systems. In this study, we develop and evaluate an operationally oriented DL framework for precipitation nowcasting that integrates multi-source data including high-resolution radar and satellite observations and automatic weather station measurements via the qPrec system over Slovakia. By incorporating satellite-derived forcing, the framework accounts for convection initiation and cloud development stage, providing a physical advantage over both classical extrapolation and radar-only deep learning methods. The framework leverages modern DL architectures, including convolutional encoder-decoder models such as U-Net and spatio-temporal transformer-based models (e.g., Earthformer), to learn the temporal evolution of precipitation fields inputs. The use of transformer-based models allows the network to capture long-range spatial dependencies and complex motion patterns that traditional CNNs may miss.The proposed models generate precipitation forecasts at a spatial resolution of 1 km and a temporal resolution of 5 minutes, with lead times of up to 60 minutes. In addition to instantaneous precipitation estimates, the framework produces 15-minute accumulated precipitation for horizons up to 120 minutes. Unlike traditional methods where predictability skill remains static across resolutions, our DL approach leverages varied spatial representations to enhance predictability at these coarser temporal scales, optimizing the forecast for different hydrological requirements. These accumulated fields can be directly applied to flash-flood hazard assessment, enabling estimation of flood likelihood as a function of rainfall intensity and duration. Model performance is evaluated using standard verification metrics such as the Fractions Skill Score, and continuous ranked probability score (reducing to MAE on deterministic outputs), showing improvement over conventional radar extrapolation methods. This study demonstrates that modern DL approaches, particularly when combined with high-resolution radar observations, offer a promising path toward next-generation operational nowcasting.