B
Beilun Wang
Researcher at University of Virginia
Publications - 36
Citations - 432
Beilun Wang is an academic researcher from University of Virginia. The author has contributed to research in topics: Graphical model & Gaussian. The author has an hindex of 8, co-authored 29 publications receiving 340 citations. Previous affiliations of Beilun Wang include Southeast University.
Papers
More filters
Proceedings ArticleDOI
Deep motif dashboard: visualizing and understanding genomic sequences using deep neural networks.
TL;DR: A toolkit called the Deep Motif Dashboard (DeMo Dashboard) is proposed which provides a suite of visualization strategies to extract motifs, or sequence patterns from deep neural network models for TFBS classification, and results indicate that a convolutional-recurrent architecture performs the best among the three architectures.
Proceedings Article
DeepCloak: Masking Deep Neural Network Models for Robustness Against Adversarial Samples
TL;DR: DeepCloak as mentioned in this paper identifies and removes unnecessary features in a DNN model to limit the capacity an attacker can use to generate adversarial samples and therefore increase the robustness against such inputs.
Posted Content
DeepCloak: Masking Deep Neural Network Models for Robustness Against Adversarial Samples
TL;DR: DeepCloak as mentioned in this paper identifies and removes unnecessary features in a DNN model to limit the capacity an attacker can use to generate adversarial samples and therefore increase the robustness against such inputs.
Posted Content
A Theoretical Framework for Robustness of (Deep) Classifiers against Adversarial Examples
Beilun Wang,Ji Gao,Yanjun Qi +2 more
TL;DR: This paper investigates the topological relationship between two (pseudo)metric spaces corresponding to predictor and oracle and develops necessary and sufficient conditions that can determine if a classifier is always robust (strong-robust) against adversarial examples according to f_2.
Posted Content
Deep Motif Dashboard: Visualizing and Understanding Genomic Sequences Using Deep Neural Networks
TL;DR: Deep Motif Dashboard (DeMo Dashboard) as mentioned in this paper provides a suite of visualization strategies to extract motifs, or sequence patterns from deep neural network models for transcription factor binding (TFBS) site classification.