Filters








45 Hits in 6.7 sec

Point Cloud Denoising via Momentum Ascent in Gradient Fields [article]

Yaping Zhao, Haitian Zheng, Zhongrui Wang, Jiebo Luo, Edmund Y. Lam
2022 arXiv   pre-print
To address these issues, we develop the momentum gradient ascent method that leverages the information of previous iterations in determining the trajectories of the points, thus improving the stability  ...  Recently, the gradient-based method was proposed to estimate the gradient fields from the noisy point clouds using neural networks, and refine the position of each point according to the estimated gradient  ...  CONCLUSION In this paper, we propose point cloud denoising via momentum ascent in gradient fields.  ... 
arXiv:2202.10094v2 fatcat:lkw3yi6wqfehdmkltcw4xstd6y

Exploring the Devil in Graph Spectral Domain for 3D Point Cloud Attacks [article]

Qianjiang Hu, Daizong Liu, Wei Hu
2022 arXiv   pre-print
The gradient field is the gradient of the log-probability function of the noisy point cloud, based on which we perform gradient ascent so as to converge each point to the underlying clean surface.  ...  In this paper, we propose a novel gradient-based dynamic point cloud denoising method, exploiting the temporal correspondence for the estimation of gradient fields – also a fundamental problem in dynamic  ...  Spectral Methods for Point Clouds In the graph spectral domain, the rough shape of a point cloud will be encoded into low-frequency components, which is suitable for denoising point clouds.  ... 
arXiv:2202.07261v2 fatcat:j4dce7nvsfhqvfb5br3xpuasiy

Crystal Diffusion Variational Autoencoder for Periodic Material Generation [article]

Tian Xie, Xiang Fu, Octavian-Eugen Ganea, Regina Barzilay, Tommi Jaakkola
2022 arXiv   pre-print
This task is difficult because stable materials only exist in a low-dimensional subspace of all possible periodic arrangements of atoms: 1) the coordinates must lie in the local energy minimum defined  ...  We significantly outperform past methods in three tasks: 1) reconstructing the input structure, 2) generating valid, diverse, and realistic materials, and 3) generating materials that optimize a specific  ...  The gradient field s X ( M |z) is used to update atom coordinates in Langevin dynamics via the force term, α j s X,t .  ... 
arXiv:2110.06197v3 fatcat:eso6lntjejbxphyro52fhojmfq

SoK: A Modularized Approach to Study the Security of Automatic Speech Recognition Systems [article]

Yuxuan Chen, Jiangshan Zhang, Xuejing Yuan, Shengzhi Zhang, Kai Chen, Xiaofeng Wang, Shanqing Guo
2021 arXiv   pre-print
where we stand in the former.  ...  More importantly, we align the research in this domain with that on security in Image Recognition System (IRS), which has been extensively studied, using the domain knowledge in the latter to help understand  ...  In [98] , the authors propose Curls-Whey black-box attack to combine gradient ascent and descent directions together, hoping to escape overfitting direction and further find more possible solutions to  ... 
arXiv:2103.10651v2 fatcat:ryllxp63hvgoxm5d6ef7n7l55a

Comparing remotely sensed observations of clouds and aerosols in the Southern Ocean with climate model simulations

Peter Kuma, Adrian McDonald, Olaf Morgenstern
2020 Zenodo  
These biases have been attributed to deficiencies in the representation of clouds during the austral summer months, either due to cloud cover or cloud albedo being too low.  ...  We modify an existing satellite lidar simulator present in the Cloud Feedback Model Intercomparison Project (CFMIP) Observational Simulator Package (COSP) for use with the ground-based lidars used in our  ...  We would like to acknowledge the financial support that made this work possible provided by the Deep South National Science Challenge via the "Clouds and Aerosols" project.  ... 
doi:10.5281/zenodo.3865850 fatcat:55x3hexvnra2vf2cms4zijlhxy

A trans-disciplinary review of deep learning research and its relevance for water resources scientists

Chaopeng Shen
2018 Water Resources Research  
Vast opportunities exist for DL to propel advances in water sciences.  ...  Deep learning (DL), a new-generation of artificial neural network research, has transformed industries, daily lives and various scientific disciplines in recent years.  ...  Gradient ascent was used to solve the maximization problem.  ... 
doi:10.1029/2018wr022643 fatcat:ruopsnchg5eg5hsiccyadinf54

Comparing remotely sensed observations of clouds and aerosols in the Southern Ocean with climate model simulations

Peter Kuma, Adrian McDonald, Olaf Morgenstern
2020 Zenodo  
These biases have been attributed to deficiencies in the representation of clouds during the austral summer months, either due to cloud cover or cloud albedo being too low.  ...  We modify an existing satellite lidar simulator present in the Cloud Feedback Model Intercomparison Project (CFMIP) Observational Simulator Package (COSP) for use with the ground-based lidars used in our  ...  We would like to acknowledge the financial support that made this work possible provided by the Deep South National Science Challenge via the 'Clouds and Aerosols' project.  ... 
doi:10.5281/zenodo.4281575 fatcat:gimsbwrfu5hv7mzswn3p5gpta4

Robust, Deep, and Reinforcement Learning for Management of Communication and Power Networks [article]

Alireza Sadeghi
2022 arXiv   pre-print
We then introduce distributionally robust learning frameworks to minimize the worst-case expected loss over a prescribed ambiguity set of training distributions quantified via Wasserstein distance.  ...  an oracle to solve the convex sub-problems; ii) stochastic proximal gradient descent-ascent, which approximates the solution of the convex sub-problems via a single gradient ascent step; and, iii) a distributionally  ...  The first algorithm relies on an -accurate maximum-oracle to solve the inner convex subproblem, while the second approximates its solution via a single gradient ascent step.  ... 
arXiv:2202.05395v1 fatcat:5v3awpoiizaxjjam5deq6yadpa

A State-of-the-Art Survey on Deep Learning Theory and Architectures

Md Zahangir Alom, Tarek M. Taha, Chris Yakopcic, Stefan Westberg, Paheding Sidike, Mst Shamima Nasrin, Mahmudul Hasan, Brian C. Van Essen, Abdul A. S. Awwal, Vijayan K. Asari
2019 Electronics  
Experimental results show state-of-the-art performance using deep learning when compared to traditional machine learning approaches in the fields of image processing, computer vision, speech recognition  ...  In recent years, deep learning has garnered tremendous success in a variety of application domains.  ...  Acknowledgments: We would like to thank all authors mentioned in the reference of this paper from whom we have learned a lot and thus made this review paper possible.  ... 
doi:10.3390/electronics8030292 fatcat:2i64q7g6kjbjvfalvzwgiggnyq

The History Began from AlexNet: A Comprehensive Survey on Deep Learning Approaches [article]

Md Zahangir Alom, Tarek M. Taha, Christopher Yakopcic, Stefan Westberg, Paheding Sidike, Mst Shamima Nasrin, Brian C Van Esesn, Abdul A S. Awwal, Vijayan K. Asari
2018 arXiv   pre-print
This new field of machine learning has been growing rapidly and applied in most of the application domains with some new modalities of applications, which helps to open new opportunity.  ...  The experimental results show state-of-the-art performance of deep learning over traditional machine learning approaches in the field of Image Processing, Computer Vision, Speech Recognition, Machine Translation  ...  RL can be applied in a different scope of field including fundamental Sciences for decision making, Machine learning from a computer science point of view, in the field of engineering and mathematics,  ... 
arXiv:1803.01164v2 fatcat:eo353y77tvckbdjcfexpaadeh4

Reports about 8 selected benchmark cases of model hierarchies

Naomi Auer, PatriciaI Barral, Jean-David Benamou, Andreas Baermann, Andreas Binder, Daniel Fernández Comesaña, Michele Girfoglio, Lena Hauberg-Lotte, Michael Hintermüller, Wilbert Ijzerman, Onkar Jadhav, Karl Knall (+18 others)
2020 Zenodo  
These will be equipped with publically available data and will be used for training in modelling, model testing, reduced order modelling, error estimation, efficiency optimization in algorithmic approaches  ...  The present document has been structured in three main parts to distinguish those contributions which are focused on coupling methods, model order reduction methods, and optimization methods.  ...  (selection field), which is given by a gradient field, generates a field free point (FFP) (or alternatively a field free line (FFL) [16] ).  ... 
doi:10.5281/zenodo.3888124 fatcat:voi4sve7jbdctctgkciykv6fz4

Advances in adversarial attacks and defenses in computer vision: A survey [article]

Naveed Akhtar, Ajmal Mian, Navid Kardan, Mubarak Shah
2021 arXiv   pre-print
Deep Learning (DL) is the most widely used tool in the contemporary field of computer vision.  ...  Since the discovery of this phenomenon in 2013~[1], it has attracted significant attention of researchers from multiple sub-fields of machine intelligence.  ...  [179] proposed a label guided GAN-based method for targeted attack on 3D point clouds in real-time.  ... 
arXiv:2108.00401v2 fatcat:23gw74oj6bblnpbpeacpg3hq5y

A Survey of Techniques All Classifiers Can Learn from Deep Networks: Models, Optimizations, and Regularization [article]

Alireza Ghods, Diane J Cook
2019 arXiv   pre-print
Non-network classifiers can employ many components found in deep neural network architectures.  ...  In this paper, we review the feature learning, optimization, and regularization methods that form a core of deep network technologies.  ...  Wiering for their valuable assistance in providing code and additional implementation details of the algorithms that were evaluated in this paper.  ... 
arXiv:1909.04791v2 fatcat:ponbnbog7rg5zltspwiawuznzy

Modern applications of machine learning in quantum sciences [article]

Anna Dawid, Julian Arnold, Borja Requena, Alexander Gresch, Marcin Płodzień, Kaelan Donatella, Kim Nicoli, Paolo Stornati, Rouven Koch, Miriam Büttner, Robert Okuła, Gorka Muñoz-Gil (+17 others)
2022 arXiv   pre-print
In these Lecture Notes, we provide a comprehensive introduction to the most recent advances in the application of machine learning methods in quantum sciences.  ...  We cover the use of deep learning and kernel methods in supervised, unsupervised, and reinforcement learning algorithms for phase classification, representation of many-body quantum states, quantum feedback  ...  We find θ * via an iterative update rule in which we estimate the gradient ∇ θ E[G|π θ ] and perform a gradient ascent step in its direction.  ... 
arXiv:2204.04198v1 fatcat:rae77aetd5hahnovchru6kjbcy

Adversarial Machine Learning Attacks and Defense Methods in the Cyber Security Domain [article]

Ihai Rosenberg and Asaf Shabtai and Yuval Elovici and Lior Rokach
2021 arXiv   pre-print
In recent years machine learning algorithms, and more specifically deep learning algorithms, have been widely used in many fields, including cyber security.  ...  Finally, we highlight some characteristics identified in recent research and discuss the impact of recent advancements in other adversarial learning domains on future research directions in the cyber security  ...  [16] suggested solving Equation 1 using gradient ascent.  ... 
arXiv:2007.02407v3 fatcat:rj3qomvg4bfb5p3atsct4winji
« Previous Showing results 1 — 15 out of 45 results