J
Jean Utke
Researcher at Argonne National Laboratory
Publications - 56
Citations - 1755
Jean Utke is an academic researcher from Argonne National Laboratory. The author has contributed to research in topics: Automatic differentiation & Computer science. The author has an hindex of 15, co-authored 47 publications receiving 1676 citations. Previous affiliations of Jean Utke include University of Chicago & Dresden University of Technology.
Papers
More filters
Book ChapterDOI
Combining Automatic Differentiation Methods for High-Dimensional Nonlinear Models
TL;DR: A recently developed hybrid approach is employed to combine reverse-mode automatic differentiation to calculate first-order derivatives and perform the required reduction in the input data stream, followed by forward-modeautomatic differentiation to Calculate higher- order derivatives with respect only to the reduced input variables.
Journal ArticleDOI
Gradient-Boosted Based Structured and Unstructured Learning
TL;DR: In this paper , the authors propose two frameworks to deal with problem settings in which both structured and unstructured data are available, which allow joint learning on both kinds of data by integrating the paradigms of boosting models and deep neural networks.
Proceedings ArticleDOI
Nested Multi-view Image Classification
TL;DR: In this paper , a nested multi-view deep network is proposed to organize instances into relevant groups that are treated differently, which is applicable to general data instances, not just images.
Journal ArticleDOI
Tricks and Plugins to GBM on Images and Sequences
TL;DR: A new algorithm for boosting Deep Convolutional Neural Networks (BoostCNN) to combine the merits of dynamic feature selection and BoostCNN, and a set of algorithms to incorporate boosting weights into a deep learning architecture based on a least squares objective function.
Journal ArticleDOI
Multi-Layer Attention-Based Explainability via Transformers for Tabular Data
TL;DR: In this paper , a graph-oriented attention-based explainability method for tabular data is proposed, where the attention matrices of all layers as a whole are mapped to a graph structure where groups of features correspond to nodes and attention values to arcs.