Author name cluster

Elke Rundensteiner

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

7 papers

1 author row

AAAI Conference 2026 Conference Paper

Neural Tangent Kernels Under Stochastic Data Augmentation

Joshua DeOliveira
Sajal Chakroborty
Walter Gerych
Elke Rundensteiner

The learning dynamics of modern neural networks remain an open problem in deep learning. The Neural Tangent Kernel (NTK) offers an elegant description of training dynamics in the infinite‑width limit, yet its classical formulation assumes a static data set. Modern model training practice departs from this strong assumption through the use of on‑the‑fly data augmentations (e.g. additive noise). In this work, we conduct an NTK-driven analysis of how data transformations affect a neural net's evolution in the function space. Our theoretical contributions characterize how repeated Gaussian perturbations from NTK-derived covariances can steer neural-net optimizations toward user‑specified behavior. These theoretical insights are empirically validated by controlled experiments. Taken together, our results lay the foundation for a promising future research direction that transforms the NTK from a descriptive to a prescriptive tool, enabling control of neural net training trajectories and behavior of inference generalization with grounded interventions.

PDF Details DOI

TMLR Journal 2025 Journal Article

Mastering SAM Prompts: A Large-Scale Empirical Study in Segmentation Refinement for Scientific Imaging

Stephen Price
Elke Rundensteiner
Danielle L. Cote

Segment Anything Model (SAM) has emerged as a prevalent tool empowering advances in vision tasks from instance segmentation, panoptic segmentation, to interactive segmentation. Leveraging powerful zero-shot capabilities enabled by visual prompts such as masks placed on the image, SAM has been shown to significantly improve tasks. Yet, a poor prompt can worsen SAM performance, risking consequences such as misdiagnoses, autonomous driving failures, or manufacturing defects. However, recent studies on visual SAM prompting remain limited, cover only a small fraction of potential prompt configurations, adopt ad-hoc evaluation strategies, and come with limited or even no rigorous analysis of the statistical significance of prompt configurations. To address this gap, we undertake the first large-scale empirical study comprehensively evaluating the impact of SAM prompt configurations on segmentation refinement. This includes 2,688 prompt configurations, including points, boxes, and masks with diverse augmentations, on four initial segmentation models for a total of 10,752 evaluations. From these results, we draw statistically significant insights along with practical guidelines for prompt design on scientific images. In particular, we recommend including a bounding box, which raised AP@50-95 by 0.320 and advise against using a coarse mask, which lowers AP@50-95 by -0.133 across all four models on microscopy data sets. We showcase that our recommended prompt configuration enables SAM to outperform leading refinement methods on multiple scientific benchmark datasets.

PDF Details

AAAI Conference 2025 Conference Paper

The Surprising Effectiveness of Infinite-Width NTKs for Characterizing and Improving Model Training

Joshua DeOliveira
Walter Gerych
Elke Rundensteiner

Developments in deep neural nets have trended towards increasingly larger overparameterized architectures, resulting in lengthy training sessions with ever more elusive training dynamics. Thus, ensuring these models learn accurate generalizable representations of data efficiently is challenging. Previous works have developed specialized techniques from data-pruning, architecture selection, pseudo-label generation, bias identification, or label refurbishment to improve downstream training. Problematically, most methods require prohibitively expensive iterative model training. In this paper, we demonstrate that we can exploit the recent neural tangent kernel (NTK) theory for understanding and improving model training behavior before ever training a model. First, we show a powerful signal derived from the NTK theory can be computed remarkably fast. We then leverage this signal for the design of a unified suite of surprisingly effective tools for the four important tasks of architecture selection, pseudo-label verification, bias identification, and label refurbishment, all requiring zero model training.

PDF Details DOI

AAAI Conference 2024 Conference Paper

Amalgamating Multi-Task Models with Heterogeneous Architectures

Jidapa Thadajarassiri
Walter Gerych
Xiangnan Kong
Elke Rundensteiner

Multi-task learning (MTL) is essential for real-world applications that handle multiple tasks simultaneously, such as selfdriving cars. MTL methods improve the performance of all tasks by utilizing information across tasks to learn a robust shared representation. However, acquiring sufficient labeled data tends to be extremely expensive, especially when having to support many tasks. Recently, Knowledge Amalgamation (KA) has emerged as an effective strategy for addressing the lack of labels by instead learning directly from pretrained models (teachers). KA learns one unified multi-task student that masters all tasks across all teachers. Existing KA for MTL works are limited to teachers with identical architectures, and thus propose layer-to-layer based approaches. Unfortunately, in practice, teachers may have heterogeneous architectures; their layers may not be aligned and their dimensionalities or scales may be incompatible. Amalgamating multi-task teachers with heterogeneous architectures remains an open problem. For this, we design Versatile Common Feature Consolidator (VENUS), the first solution to this problem. VENUS fuses knowledge from the shared representations of each teacher into one unified generalized representation for all tasks. Specifically, we design the Feature Consolidator network that leverages an array of teacher-specific trainable adaptors. These adaptors enable the student to learn from multiple teachers, even if they have incompatible learned representations. We demonstrate that VENUS outperforms five alternative methods on numerous benchmark datasets across a broad spectrum of experiments.

PDF Details DOI

AAAI Conference 2023 Conference Paper

Knowledge Amalgamation for Multi-Label Classification via Label Dependency Transfer

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.

PDF Details DOI

AAAI Conference 2022 Conference Paper

Recovering the Propensity Score from Biased Positive Unlabeled Data

Walter Gerych
Thomas Hartvigsen
Luke Buquicchio
Emmanuel Agu
Elke Rundensteiner

Positive-Unlabeled (PU) learning methods train a classifier to distinguish between the positive and negative classes given only positive and unlabeled data. While traditional PU methods require the labeled positive samples to be an unbiased sample of the positive distribution, in practice the labeled sample is often a biased draw from the true distribution. Prior work shows that if we know the likelihood that each positive instance will be selected for labeling, referred to as the propensity score, then the biased sample can be used for PU learning. Unfortunately, no prior work has been proposed an inference strategy for which the propensity score is identifiable. In this work, we propose two sets of assumptions under which the propensity score can be uniquely determined: one in which no assumption is made on the functional form of the propensity score (requiring assumptions on the data distribution), and the second which loosens the data assumptions while assuming a functional form for the propensity score. We then propose inference strategies for each case. Our empirical study shows that our approach significantly outperforms the state-of-the-art propensity estimation methods on a rich variety of benchmark datasets.

PDF Details

JBHI Journal 2020 Journal Article

Screening For Depression With Retrospectively Harvested Private Versus Public Text

ML Tlachac
Elke Rundensteiner

Depression is the leading cause of disability, often undiagnosed, and one of the most treatable mood disorders. As such, unobtrusively diagnosing depression is important. Many studies are starting to utilize machine learning for depression sensing from social media and Smartphone data to replace the survey instruments currently employed to screen for depression. In this study, we compare the ability of a privately versus a publicly available modality to screen for depression. Specifically, we leverage between two weeks and a year of text messages and tweets to predict scores from the Patient Health Questionnaire-9, a prevalent depression screening instrument. This is the first study to leverage the retrospectively-harvested crowd-sourced texts and tweets within the combined Moodable and EMU datasets. Our approach involves comprehensive feature engineering, feature selection, and machine learning. Our 245 features encompass word category frequencies, part of speech tag frequencies, sentiment, and volume. The best model is Logistic Regression built on the top ten features from two weeks of text data. This model achieves an average F1 score of 0. 806, AUC of 0. 832, and recall of 0. 925. We discuss the implications of the selected features, temporal quantity of data, and modality.

Details DOI