Path Augmentation

Path augmentation is a technique that modifies or enhances the pathways used in various computational processes, aiming to improve performance or robustness. Current research focuses on applications ranging from improving the transferability of adversarial examples in deep learning to enhancing the inference of stochastic systems and optimizing pathfinding algorithms in graph networks. These methods often involve augmenting data or modifying network weights to learn more effective paths, leading to improvements in accuracy, efficiency, and generalization across diverse domains. The impact of path augmentation is significant, offering potential for advancements in areas such as computer vision, system identification, and artificial intelligence.

Papers

March 28, 2023

Improving the Transferability of Adversarial Samples by Path-Augmented Method
Jianping Zhang, Jen-tse Huang, Wenxuan Wang, Yichen Li, Weibin Wu, Xiaosen Wang, Yuxin Su, Michael R. Lyu
Task Transferability Adversarial Sample Adversarial Noise Adversarial Data Collection Gradient Path Path Augmentation

January 19, 2023

Geometric path augmentation for inference of sparsely observed stochastic nonlinear systems
Dimitra Maoutsa
Scientific Inference Semi Parametric Interacting System Stochastic Nonlinear System Random Force Path Augmentation

April 11, 2022

Consistency Learning via Decoding Path Augmentation for Transformers in Human Object Interaction Detection
Jihwan Park, SeungJun Lee, Hwan Heo, Hyeong Kyu Choi, Hyunwoo J. Kim
Transformer Megatron Decepticons Human Object Interaction Detection Consistency Learning Hoi Detection Task Consistency Holistic Scene Understanding Path Augmentation

January 26, 2022

Combining Optimal Path Search With Task-Dependent Learning in a Neural Network
Tomas Kulvicius, Minija Tamosiunaite, Florentin Wörgötter
Neural Network Task Specific Optimal Path Path Finding Network Learning Bellman Update Path Augmentation

Path Augmentation

Papers

Improving the Transferability of Adversarial Samples by Path-Augmented Method

Geometric path augmentation for inference of sparsely observed stochastic nonlinear systems

Consistency Learning via Decoding Path Augmentation for Transformers in Human Object Interaction Detection

Combining Optimal Path Search With Task-Dependent Learning in a Neural Network