WGWalter Gerych

Walter Gerych

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Walter Gerych has a total of 14 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    An Investigation of Memorization Risk in Healthcare Foundation Models

    by Sana Tonekaboni, Lena Stempfle, Adibvafa Fallahpour, Walter Gerych, Marzyeh Ghassemi

    Foundation models trained on large-scale de-identified electronic health records (EHRs) hold promise for clinical applications. However, their capacity to memorize patient information raises important privacy concerns. In this work, we introduce a suite of black-box evaluation tests to assess privacy-related memorization risks in foundation models trained on structured EHR data. Our framework includes methods for probing memorization at both the embedding and generative levels, and aims to distinguish between model generalization and harmful memorization in clinically relevant settings. We contextualize memorization in terms of its potential to compromise patient privacy, particularly for vulnerable subgroups. We validate our approach on a publicly available EHR foundation model and release an open-source toolkit to facilitate reproducible and collaborative privacy assessments in healthcare AI.

  • Open Access

    Knowledge Amalgamation for Multi-Label Classification via Label Dependency Transfer

    by Jidapa Thadajarassiri, Thomas Hartvigsen, Walter Gerych, Xiangnan Kong, Elke Rundensteiner

    Multi-label classification (MLC), which assigns multiple labels to each instance, is crucial to domains from computer vision to text mining. Conventional methods for MLC require huge amounts of labeled data to capture complex dependencies between labels. However, such labeled datasets are expensive, or even impossible, to acquire. Worse yet, these pre-trained MLC models can only be used for the particular label set covered in the training data. Despite this severe limitation, few methods exist for expanding the set of labels predicted by pre-trained models. Instead, we acquire vast amounts of new labeled data and retrain a new model from scratch. Here, we propose combining the knowledge from multiple pre-trained models (teachers) to train a new student model that covers the union of the labels predicted by this set of teachers. This student supports a broader label set than any one of its teachers without using labeled data. We call this new problem knowledge amalgamation for multi-label classification. Our new method, Adaptive KNowledge Transfer (ANT), trains a student by learning from each teacher’s partial knowledge of label dependencies to infer the global dependencies between all labels across the teachers. We show that ANT succeeds in unifying label dependencies among teachers, outperforming five state-of-the-art methods on eight real-world datasets.


  • Open Access

    Domain Adaptation Methods for Lab-to-Field Human Context Recognition

    by Abdulaziz Alajaji, Walter Gerych, Luke Buquicchio, Kavin Chandrasekaran, Hamid Mansoor, Emmanuel Agu, Elke Rundensteiner

    Human context recognition (HCR) using sensor data is a crucial task in Context-Aware (CA) applications in domains such as healthcare and security. Supervised machine learning HCR models are trained using smartphone HCR datasets that are scripted or gathered in-the-wild. Scripted datasets are most accurate because of their consistent visit patterns. Supervised machine learning HCR models perform well on scripted datasets but poorly on realistic data. In-the-wild datasets are more realistic, but cause HCR models to perform worse due to data imbalance, missing or incorrect labels, and a wide variety of phone placements and device types. Lab-to-field approaches learn a robust data representation from a scripted, high-fidelity dataset, which is then used for enhancing performance on a noisy, in-the-wild dataset with similar labels. This research introduces Triplet-based Domain Adaptation for Context REcognition (Triple-DARE), a lab-to-field neural network method that combines three unique loss functions to enhance intra-class compactness and inter-class separation within the embedding space of multi-labeled datasets: (1) domain alignment loss in order to learn domain-invariant embeddings; (2) classification loss to preserve task-discriminative features; and (3) joint fusion triplet loss. Rigorous evaluations showed that Triple-DARE achieved 6.3% and 4.5% higher F1-score and classification, respectively, than state-of-the-art HCR baselines and outperformed non-adaptive HCR models by 44.6% and 10.7%, respectively.


  • Exploratory Data Analysis of Population Level Smartphone-Sensed Data

    by Hamid Mansoor, Walter Gerych, Abdulaziz Alajaji, Luke Buquicchio, Kavin Chandrasekaran, Emmanuel Agu, Elke Rundensteiner

  • Visual Analytics of Smartphone-Sensed Human Behavior and Health

    by Hamid Mansoor, Walter Gerych, Abdulaziz Alajaji, Luke Buquicchio, Kavin Chandrasekaran, Emmanuel Agu, Elke A. Rundensteiner

    Smartphone health sensing tools, which analyze passively gathered human behavior data, can provide clinicians with a longitudinal view of their patients’ ailments in natural settings. In this Visualization Viewpoints article, we postulate that interactive visual analytics (IVA) can assist data scientists during the development of such tools by facilitating the discovery and correction of wrong or missing user-provided ground-truth health annotations. IVA can also assist clinicians in making sense of their patients’ behaviors by providing additional contextual and semantic information. We review the current state-of-the-art, outline unique challenges, and illustrate our viewpoints using our work as well as those of other researchers. Finally, we articulate open challenges in this exciting and emerging field of research.


  • Smartphone Health Biomarkers: Positive Unlabeled Learning of In-the-Wild Contexts

    by Abdulaziz Alajaji, Walter Gerych, Luke Buquicchio, Kavin Chandrasekaran, Hamid Mansoor, Emmanuel Agu, Elke A. Rundensteiner

    There has recently been increased interest in context-aware mobile sensing applications due to the ubiquity of sensor-rich smartphones. Our DARPA-funded Warfighter Analytics for Smartphone Healthcare (WASH) project is exploring passive assessment methods using smartphone biomarkers and context-specific tests. Our envisioned context-specific assessments require accurate recognition of specific smartphone user contexts. Existing context datasets were either scripted or in-the-wild. Scripted datasets have accurate context labels but user behaviors are not realistic. In-the-wild datasets have realistic user behaviors but often have wrong or missing labels. We introduce a novel coincident data gathering study design in which data were gathered for the same contexts using both a scripted and in-the-wild study. We then propose positive unlabeled context learning (PUCL), a transductive method to transfer knowledge from highly accurate labels of the scripted dataset to the less accurate in-the-wild dataset. Our PUCL approach for context recognition outperforms state-of-the-art methods with an increase of over 3% in balanced accuracy.