Filters








22,656 Hits in 2.7 sec

Adversarial Memory Networks for Action Prediction [article]

Zhiqiang Tao, Yue Bai, Handong Zhao, Sheng Li, Yu Kong, Yun Fu
2021 arXiv   pre-print
In this study, we propose adversarial memory networks (AMemNet) to generate the "full video" feature conditioning on a partial video query from two new aspects.  ...  Action prediction aims to infer the forthcoming human action with partially-observed videos, which is a challenging task due to the limited information underlying early observations.  ...  Conclusion In this paper, we presented a novel two-stream adversarial memory networks (AMemNet) model for the action prediction task.  ... 
arXiv:2112.09875v1 fatcat:jhksvwsnxzfhvcrxlzzyroiyki

Table of Contents

2019 2019 16th Conference on Computer and Robot Vision (CRV)  
Human Action Prediction STAR-Net: Action Recognition using Spatio-Temporal Activation Reprojection 49 William McNally (University of Waterloo), Alexander Wong (University of Waterloo), and John McPhee  ...  Abnormal Trajectory Classifier 65 Pankaj Roy (Polytechnique Montréal -LITIV) and Guillaume-Alexandre Bilodeau (Polytechnique Montréal -LITIV) Hierarchically-Fused Generative Adversarial Network for Text  ... 
doi:10.1109/crv.2019.00004 fatcat:svtlhabybfeibjz3ub57woazcu

Task Specific Visual Saliency Prediction with Memory Augmented Conditional Generative Adversarial Networks [article]

Tharindu Fernando, Simon Denman, Sridha Sridharan, Clinton Fookes
2018 arXiv   pre-print
To address this limitation, we propose a novel saliency estimation model which leverages the semantic modelling power of conditional generative adversarial networks together with memory architectures which  ...  Our studies not only shed light on a novel application area for generative adversarial networks, but also emphasise the importance of task specific saliency modelling and demonstrate the plausibility of  ...  Generative Adversarial Networks Generative adversarial networks (GAN), which belong to the family of generative models, have achieved promising results for pixel-to-pixel synthesis [41] .  ... 
arXiv:1803.03354v1 fatcat:xtcc4agytndphaqcj3mbqvfmdu

Task Specific Visual Saliency Prediction with Memory Augmented Conditional Generative Adversarial Networks

Tharindu Fernando, Simon Denman, Sridha Sridharan, Clinton Fookes
2018 2018 IEEE Winter Conference on Applications of Computer Vision (WACV)  
Generative Adversarial Networks Generative adversarial networks (GAN), which belong to the family of generative models, have achieved promising results for pixel-to-pixel synthesis [41] .  ...  We draw our inspiration from the recent success of Generative Adversarial Networks (GAN) [25] [26] [27] [28] [29] for pixel to pixel translation tasks.  ... 
doi:10.1109/wacv.2018.00172 dblp:conf/wacv/FernandoDSF18a fatcat:od5jfo6a4raczl3ly47qb5in2u

Generating the Future with Adversarial Transformers

Carl Vondrick, Antonio Torralba
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Our approach explicitly disentangles the model's memory from the prediction, which helps the model learn desirable invariances.  ...  We believe predictive models have many applications in robotics, health-care, and video understanding.  ...  Acknowledgements: Funding for this work was partially provided by the Google PhD fellowship to CV. We acknowledge NVidia Corporation for hardware donations.  ... 
doi:10.1109/cvpr.2017.319 dblp:conf/cvpr/Vondrick017 fatcat:xdhcp5tg3nct7odsu3f2dz7uia

DADI: Dynamic Discovery of Fair Information with Adversarial Reinforcement Learning [article]

Michiel A. Bakker, Duy Patrick Tu, Humberto Riverón Valdés, Krishna P. Gummadi, Kush R. Varshney, Adrian Weller, Alex Pentland
2019 arXiv   pre-print
We introduce a framework for dynamic adversarial discovery of information (DADI), motivated by a scenario where information (a feature set) is used by third parties with unknown objectives.  ...  Finally, we demonstrate empirically, using two real-world datasets, that we can trade-off fairness and predictive performance  ...  Acknowledgements The authors want to thank IBM for access to their compute resources and Prasanna Sattigeri for helpful discussions.  ... 
arXiv:1910.13983v1 fatcat:5hgkx32fxbenrk3sbxslqgxwcu

A multi-task learning based hybrid prediction algorithm for privacy preserving human activity recognition framework

Vijaya Kumar Kambala, Harikiran Jonnadula
2021 Bulletin of Electrical Engineering and Informatics  
In this paper, we proposed a multi-task learning based hybrid prediction algorithm (MTL-HPA) towards realising privacy preserving human activity recognition framework (PPHARF).  ...  There is need for protecting privacy of people while videos are used purposefully based on objective functions. One such use case is human activity recognition without disclosing human identity.  ...  Action recognition The action recognition for face anonymized image is done using bi-directional long short term memory (BLSTM) network. long short term memory (LSTM) is an extension of recurrent neural  ... 
doi:10.11591/eei.v10i6.3204 fatcat:xs64vmm7urhjdighpc23pmtgky

Adversarial Refinement Network for Human Motion Prediction [article]

Xianjin Chao, Yanrui Bin, Wenqing Chu, Xuan Cao, Yanhao Ge, Chengjie Wang, Jilin Li, Feiyue Huang, Howard Leung
2020 arXiv   pre-print
Specifically, we take both the historical motion sequences and coarse prediction as input of our cascaded refinement network to predict refined human motion and strengthen the refinement network with adversarial  ...  To predict more accurate future human motion, we propose an Adversarial Refinement Network (ARNet) following a simple yet effective coarse-to-fine mechanism with novel adversarial error augmentation.  ...  networks used for human motion prediction.  ... 
arXiv:2011.11221v2 fatcat:7w6quwsgd5c4rdy2echwm4wlcq

Sparse Adversarial Perturbations for Videos [article]

Xingxing Wei, Jun Zhu, Hang Su
2018 arXiv   pre-print
We choose the action recognition as the targeted task, and networks with a CNN+RNN architecture as threat models to verify our method.  ...  To this end, we propose an l2,1-norm based optimization algorithm to compute the sparse adversarial perturbations for videos.  ...  The red dotted line is the predicted frame-level label indices for the clean video, and black dotted line is the predicted frame-level label indices for the adversarial video, both by the action recognition  ... 
arXiv:1803.02536v1 fatcat:fumttgcdfjbj3ng2p35strhqay

Network Penetration Intrusion Prediction Based on Attention Seq2seq Model

Tianxiang Yu, Yang Xin, Hongliang Zhu, Qifeng Tang, Yuling Chen, AnMin Fu
2022 Security and Communication Networks  
In order to supplement this part of research, this paper reports the prediction of network penetration intrusion sequence for the first time.  ...  Only by predicting the next possible attack can we prevent the corresponding intrusion or cheat adversary more efficiently.  ...  the real network and there is some leakage in the network. at's the main motivation for us to perform the network penetration intrusion prediction.  ... 
doi:10.1155/2022/6012232 fatcat:s3tevtlv4jcupfxhfh3tqa3nxm

Model-based actor-critic: GAN (model generator) + DRL (actor-critic) => AGI [article]

Aras Dargazany
2021 arXiv   pre-print
) environment model to the actor-critic (model-free) architecture which results in a model-based actor-critic architecture with temporal-differencing (TD) error and an episodic memory.  ...  visualization; (B) online learning (of real or simulated devices) like DRL in RL setting (with/out environment reward) such as (real or simulated) robotics and control; Our core proposal is adding an (generative/predictive  ...  The aim of AC methods is to simultaneously learn an action-value function that predicts the total expected discounted  ... 
arXiv:2004.04574v9 fatcat:jqzrwh2slnfhpdzz74qseymguy

Sparse Adversarial Perturbations for Videos

Xingxing Wei, Jun Zhu, Sha Yuan, Hang Su
2019 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
We choose the action recognition as the targeted task, and networks with a CNN+RNN architecture as threat models to verify our method.  ...  Although adversarial samples of deep neural networks (DNNs) have been intensively studied on static images, their extensions in videos are never explored.  ...  The red dotted line is the predicted frame-level label indices for the clean video, and black dotted line is the predicted frame-level label indices for the adversarial video, both by the action recognition  ... 
doi:10.1609/aaai.v33i01.33018973 fatcat:cbrwfn5c35cspp4f4rw4xpbzei

Detecting Adversarial Attacks on Neural Network Policies with Visual Foresight [article]

Yen-Chen Lin, Ming-Yu Liu, Min Sun, Jia-Bin Huang
2017 arXiv   pre-print
Our core idea is that the adversarial examples targeting at a neural network-based policy are not effective for the frame prediction model.  ...  action-conditioned frame prediction module, we can detect the presence of adversarial examples.  ...  Acknowledgements We gratefully acknowledge the support of NVIDIA Corporation with the donation of the Titan X GPU used for this research.  ... 
arXiv:1710.00814v1 fatcat:srvljepqvzg4bmo2suwot26kg4

Adversarial Generative Grammars for Human Activity Prediction [article]

AJ Piergiovanni, Anelia Angelova, Alexander Toshev, Michael S. Ryoo
2020 arXiv   pre-print
In this paper we propose an adversarial generative grammar model for future prediction.  ...  The adversarial generative grammar is evaluated on the Charades, MultiTHUMOS, Human3.6M, and 50 Salads datasets and on two activity prediction tasks: future 3D human pose prediction and future activity  ...  Our adversarial stochastic sampling process allows for much more memory-and computationally-efficient learning without such enumeration.  ... 
arXiv:2008.04888v2 fatcat:7lipm6i3ijghrghuxm3vpuvgzy

Robust Adversarial Attacks Detection based on Explainable Deep Reinforcement Learning For UAV Guidance and Planning [article]

Thomas Hickling, Nabil Aouf, Phillippa Spencer
2022 arXiv   pre-print
The second detector is developed based on a Long Short Term Memory (LSTM) network and achieves an accuracy of 91\% with much faster computing times when compared to the CNN based detector.  ...  A Realistic Synthetic environment for UAV explainable DRL based planning and guidance including obstacles and adversarial attacks is built. Two adversarial attack detectors are proposed.  ...  Lin et al.(2017) [17] use an action-conditioned frame prediction module to detect attacks.  ... 
arXiv:2206.02670v2 fatcat:yjtbtkb37vgcblvet3b3l7fuvu
« Previous Showing results 1 — 15 out of 22,656 results