2,203 Hits in 4.3 sec

Black-box Gradient Attack on Graph Neural Networks: Deeper Insights in Graph-based Attack and Defense [article]

Haoxi Zhan, Xiaobing Pei
2021 arXiv   pre-print
Secondly, to overcome the drawbacks, we propose the Black-Box Gradient Attack (BBGA) algorithm.  ...  In this paper, we develop deeper insights into the Mettack algorithm, which is a representative grey-box attacking method, and then we propose a gradient-based black-box attacking algorithm.  ...  Since the access to the training set is also prohibited, reinforcement learning is introduced to perform black-box attacks [9] .  ... 
arXiv:2104.15061v2 fatcat:a7odz5p5mjb73pzb6urkrfajcu

Known Operator Learning and Hybrid Machine Learning in Medical Imaging — A Review of the Past, the Present, and the Future [article]

Andreas Maier, Harald Köstler, Marco Heisig, Patrick Krauss, Seung Hee Yang
2021 arXiv   pre-print
Towards the end of the article, we will investigate a few future directions and point out relevant areas in which hybrid modelling, meta learning, and other domains will likely be able to drive the state-of-the-art  ...  Even domains such as physical simulation and scanner and acquisition design are being addressed using machine learning grey box modelling approaches.  ...  810316 to AM).  ... 
arXiv:2108.04543v1 fatcat:vnb7ve55bnevfbfd2fxiiojmnq

Defining a Lingua Franca to Open the Black Box of a Naïve Bayes Recommender [article]

Kenneth L. Hess, Hugo D. Paz
2017 arXiv   pre-print
Many AI systems have a black box nature that makes it difficult to understand how they make their recommendations.  ...  To penetrate our Na\"ive Bayes recommender's black box, we first asked, what do we want to know from our system, and how can it be obtained?  ...  In many cases, the darkness of a black box can be measured by how much effort it takes to learn what you want to know.  ... 
arXiv:1709.07528v1 fatcat:2micdsb6fve7npwuu4ck2tmtxi

Automatically Composing Representation Transformations as a Means for Generalization [article]

Michael B. Chang, Abhishek Gupta, Sergey Levine, Thomas L. Griffiths
2019 arXiv   pre-print
As a first step for tackling compositional generalization, we introduce the compositional recursive learner, a domain-general framework for learning algorithmic procedures for composing representation  ...  This paper introduces the compositional problem graph as a broadly applicable formalism to relate tasks of different complexity in terms of problems with shared subproblems.  ...  ACKNOWLEDGMENTS The authors would like to thank the anonymous ICLR reviewers and commenters, Alyosha Efros, Dinesh Jayaraman, Pulkit Agrawal, Jason Peng, Erin Grant, Rachit Dubey, Thanard Kurutach, Parsa  ... 
arXiv:1807.04640v2 fatcat:rupxorh2lndmpgophzwjzo4a5a

ML-misfit: Learning a robust misfit function for full-waveform inversion using machine learning [article]

Bingbing Sun, Tariq Alkhalifah
2020 arXiv   pre-print
To guarantee the resulting learned misfit is a metric, we accommodate the symmetry of the misfit with respect to its input and a Hinge loss regularization term in a meta-loss function to satisfy the "triangle  ...  In the framework of meta-learning, we train the network by running FWI to invert for randomly generated velocity models and update the parameters of the NN by minimizing the meta-loss, which is defined  ...  is introduced to guarantee symmetry of the misfit function with respect to the inputs  ... 
arXiv:2002.03163v2 fatcat:yrtuvx4gpjhplkp7nofm5j6pgu

Explainability in Deep Reinforcement Learning [article]

Alexandre Heuillet, Fabien Couthouis, Natalia Díaz-Rodríguez
2020 arXiv   pre-print
of what is still considered a black box.  ...  We review recent works in the direction to attain Explainable Reinforcement Learning (XRL), a relatively new subfield of Explainable Artificial Intelligence, intended to be used in general public applications  ...  We also would like to thank Frédéric Herbreteau and Adrien Bennetot for their help and support.  ... 
arXiv:2008.06693v4 fatcat:r62o6dabufc4ddfklhjx3lgjnq

Multi-Dimensional Deep Memory Atari-Go Players for Parameter Exploring Policy Gradients [chapter]

Mandy Grüttner, Frank Sehnke, Tom Schaul, Jürgen Schmidhuber
2010 Lecture Notes in Computer Science  
In order to improve the convergence rate, as well as the ultimate performance, we train those networks using Policy Gradients with Parameter-based Exploration, a recently developed Reinforcement Learning  ...  In this paper we attempt to tackle this challenge through a combination of two recent developments in Machine Learning.  ...  Conclusion In this paper we have introduced different methods of Machine Learning: PGPE, an algorithm based on a gradient based search through model parameter space, ES and CMA-ES, based on population  ... 
doi:10.1007/978-3-642-15822-3_14 fatcat:qbyhg4nnszaavhq3wuln3vbguq

Vision-Based Manipulators Need to Also See from Their Hands [article]

Kyle Hsu, Moo Jin Kim, Rafael Rafailov, Jiajun Wu, Chelsea Finn
2022 arXiv   pre-print
On six representative manipulation tasks with varying hand-centric observability adapted from the Meta-World benchmark, this results in a state-of-the-art reinforcement learning agent operating from both  ...  To mitigate this, we propose to regularize the third-person information stream via a variational information bottleneck.  ...  We expect this to be sufficient for independent replication of our main findings. Separately, we have included links to code used for our simulation experiments on our project website.  ... 
arXiv:2203.12677v1 fatcat:jhsgbiu2jrbkdm724wnyrpxova

Multi-label Classification for the Generation of Sub-problems in Time-constrained Combinatorial Optimization

Luca Mossina, Emmanuel Rachelson, Daniel Delahaye
2019 Proceedings of the 8th International Conference on Operations Research and Enterprise Systems  
a supervised learning model for multi-label classification.  ...  This model is exploited to predict a subset of decision variables to be set heuristically to a certain reference value, thus becoming fixed parameters in the original problem.  ...  ACKNOWLEDGEMENTS This research benefited from the support of the "FMJH Program Gaspard Monge in optimization and operation research", and from the support to this program from EDF.  ... 
doi:10.5220/0007396601330141 dblp:conf/icores/MossinaRD19 fatcat:hobegdsvvvdw3hkz2lhhqephsu

Exploring to learn visual saliency: The RL-IAC approach [article]

Celine Craye, Timothee Lesort, David Filliat, Jean-Francois Goudou
2018 arXiv   pre-print
The proposed exploration, called Reinforcement Learning-Intelligent Adaptive Curiosity (RL-IAC) is able to drive the robot's exploration so that samples selected by the robot are likely to improve the  ...  This model of saliency can also be exploited to produce bounding box proposals around objects of interest.  ...  ACKNOWLEDGMENT The authors would like to thank the INRIA Flowers team, and especially Pierre-Yves Oudeyer for the valuable help on the IAC aspect.  ... 
arXiv:1804.00435v1 fatcat:3gcttrpnmnf43pxvizxmxiy2p4

Achilles Heels for AGI/ASI via Decision Theoretic Adversaries [article]

Stephen Casper
2022 arXiv   pre-print
As progress in AI continues to advance, it is crucial to know how advanced systems will make choices and in what ways they may fail.  ...  which cause them to make obviously irrational decisions in adversarial settings.  ...  I would also like to thank the Effective Altruism and LessWrong communities for fostering a marketplace of ideas surrounding crucial questions in decision theory and AI alignment.  ... 
arXiv:2010.05418v4 fatcat:h5l2xmyxv5adtnt33t2674qxay

Knowledge Augmented Machine Learning with Applications in Autonomous Driving: A Survey [article]

Julian Wörmann, Daniel Bogdoll, Etienne Bührle, Han Chen, Evaristus Fuh Chuo, Kostadin Cvejoski, Ludger van Elst, Tobias Gleißner, Philip Gottschall, Stefan Griesche, Christian Hellert, Christian Hesels (+34 others)
2022 arXiv   pre-print
The existence of representative datasets is a prerequisite of many successful artificial intelligence and machine learning models.  ...  The identified approaches are structured according to the categories integration, extraction and conformity. Special attention is given to applications in the field of autonomous driving.  ...  The opaque black-box nature of deep learning models therefore makes it difficult to meet such demands.  ... 
arXiv:2205.04712v1 fatcat:u2bgxr2ctnfdjcdbruzrtjwot4

A Perspective on Deep Learning for Molecular Modeling and Simulations [article]

Jun Zhang, Yao-Kun Lei, Zhen Zhang, Junhan Chang, Maodong Li, Xu Han, Lijiang Yang, Yi Isaac Yang, Yi Qin Gao
2020 arXiv   pre-print
For this purpose, we summarized several representative applications, ranging from supervised to unsupervised and reinforcement learning, and discussed their connections with the emerging trends in deep  ...  learning.  ...  In other words, instead of regarding the ANN as a sheer "black box", a rational design of the architecture according to the input data is crucial to the performance.  ... 
arXiv:2004.13011v1 fatcat:xtreyq6i7bgtngwwwmc4rpc6fy

Learning to Search with MCTSnets [article]

Arthur Guez, Théophane Weber, Ioannis Antonoglou, Karen Simonyan, Oriol Vinyals, Daan Wierstra, Rémi Munos, David Silver
2018 arXiv   pre-print
In this paper we instead learn where, what and how to search.  ...  When applied to small searches in the well known planning problem Sokoban, the learned search algorithm significantly outperformed MCTS baselines.  ...  Kocsis et al. (2005) apply black-box optimisation to learn the meta-parameters controlling an alpha-beta search, but do not learn fine-grained control over the search decisions.  ... 
arXiv:1802.04697v2 fatcat:qthd5yjluzatznd7ltknndo3xu

Evolving the Materials Genome: How Machine Learning Is Fueling the Next Generation of Materials Discovery

Changwon Suh, Clyde Fare, James A. Warren, Edward O. Pyzer-Knapp
2020 Annual review of materials research (Print)  
We focus on three major themes—learning to see, learning to estimate, and learning to search materials—to show how advanced computational learning technologies are rapidly and successfully used to solve  ...  Machine learning, applied to chemical and materials data, is transforming the field of materials discovery and design, yet significant work is still required to fully take advantage of machine learning  ...  Bayesian Optimization Bayesian optimization is an efficient black-box optimization technique often used for the optimization of expensive functions.  ... 
doi:10.1146/annurev-matsci-082019-105100 fatcat:dyxljg2mu5grzlakeeatvyymd4
« Previous Showing results 1 — 15 out of 2,203 results