Filters








4,188 Hits in 8.2 sec

A Connection between Generative Adversarial Networks, Inverse Reinforcement Learning, and Energy-Based Models [article]

Chelsea Finn, Paul Christiano, Pieter Abbeel, Sergey Levine
2016 arXiv   pre-print
Generative adversarial networks (GANs) are a recently proposed class of generative models in which a generator is trained to optimize a cost function that is being simultaneously learned by a discriminator  ...  Interestingly, maximum entropy IRL is a special case of an energy-based model.  ...  Acknowledgments The authors would like to thank Ian Goodfellow and Joan Bruna for insightful discussions.  ... 
arXiv:1611.03852v3 fatcat:6igqsnwwkrbbtbfnsshkvdfjby

Connecting Generative Adversarial Networks and Actor-Critic Methods [article]

David Pfau, Oriol Vinyals
2017 arXiv   pre-print
Both generative adversarial networks (GAN) in unsupervised learning and actor-critic methods in reinforcement learning (RL) have gained a reputation for being difficult to optimize.  ...  We review the strategies for stabilizing training for each class of models, both those that generalize between the two and those that are particular to that model.  ...  It has also been shown in [35] that generative adversarial networks have a close connection with maximum entropy inverse reinforcement learning.  ... 
arXiv:1610.01945v3 fatcat:2gcikcjkvngbpe3ej3y4bm4onu

Deep Generative Models with Learnable Knowledge Constraints [article]

Zhiting Hu, Zichao Yang, Ruslan Salakhutdinov, Xiaodan Liang, Lianhui Qin, Haoye Dong, Eric Xing
2018 arXiv   pre-print
In this paper, we establish mathematical correspondence between PR and reinforcement learning (RL), and, based on the connection, expand PR to learn constraints as the extrinsic reward in RL.  ...  Experiments on human image generation and templated sentence generation show models with learned knowledge constraints by our algorithm greatly improve over base generative models.  ...  Acknowledgment This material is based upon work supported by the National Science Foundation grant IIS1563887.  ... 
arXiv:1806.09764v2 fatcat:6apd74tnlfftjpbkeyjglim3uy

Energy-Based Sequence GANs for Recommendation and Their Connection to Imitation Learning [article]

Jaeyoon Yoo, Heonseok Ha, Jihun Yi, Jongha Ryu, Chanju Kim, Jung-Woo Ha, Young-Han Kim, Sungroh Yoon
2017 arXiv   pre-print
Towards this goal, energy-based sequence generative adversarial nets (EB-SeqGANs) are adopted for recommendation by learning a generative model for the time series of user-preferred items.  ...  Recommender systems aim to find an accurate and efficient mapping from historic data of user-preferred items to a new item that is to be liked by a user.  ...  BACKGROUND 2.1 Generative Adversarial Networks Generative adversarial networks [9] is a class of generative models that learns by game theoretic competition between a generator G and a discriminator  ... 
arXiv:1706.09200v1 fatcat:j7bep3vgurcend7axrjk5i3o3m

Adversarial Resilience Learning - Towards Systemic Vulnerability Analysis for Large and Complex Systems [article]

Lars Fischer, Jan-Menno Memmen, Eric MSP Veith, Martin Tröschel
2018 arXiv   pre-print
This paper introduces Adversarial Resilience Learning (ARL), a concept to model, train, and analyze artificial neural networks as representations of competitive agents in highly complex systems.  ...  We provide the constitutive nomenclature of ARL and, based on it, the description of experimental setups and results of a preliminary implementation of ARL in simulated power systems.  ...  A modern application of unsupervised learning has emerged in the concept of Generative Adversarial Networks (GANs). Here, one network, called the generator network, creates solution candidatesi. e.  ... 
arXiv:1811.06447v1 fatcat:qapzyhrurzfflixbgzsqlbcsyy

Non-Adversarial Imitation Learning and its Connections to Adversarial Methods [article]

Oleg Arenz, Gerhard Neumann
2020 arXiv   pre-print
Many modern methods for imitation learning and inverse reinforcement learning, such as GAIL or AIRL, are based on an adversarial formulation.  ...  We address these problems by proposing a framework for non-adversarial imitation learning.  ...  Adversarial Inverse Reinforcement Learning AIRL is based on an adversarial formulation and models the reward function r θ as a neural network which is trained as a special type of discriminator.  ... 
arXiv:2008.03525v1 fatcat:hgd5t5i5hfbqjh2yhym42jmffu

Intelligent on-demand design of phononic metamaterials

Yabin Jin, Liangshu He, Zhihui Wen, Bohayra Mortazavi, Hongwei Guo, Daniel Torrent, Bahram Djafari-Rouhani, Timon Rabczuk, Xiaoying Zhuang, Yan Li
2022 Nanophotonics  
Many advanced machine learning algorithms, such as deep neural networks, unsupervised manifold clustering, reinforcement learning and so forth, have been widely and deeply investigated for structural design  ...  Machine learning provides a powerful means of achieving an efficient and accurate design process by exploring nonlinear physical patterns in high-dimensional space, based on data sets of candidate structures  ...  An inverse design network connected to a forward modeling network.  ... 
doi:10.1515/nanoph-2021-0639 fatcat:tonxwxrztvhudmgztlx2h7kwre

Reparameterized Variational Divergence Minimization for Stable Imitation [article]

Dilip Arumugam, Debadeepta Dey, Alekh Agarwal, Asli Celikyilmaz, Elnaz Nouri, Bill Dolan
2020 arXiv   pre-print
We contribute a reparameterization trick for adversarial imitation learning to alleviate the optimization challenges of the promising f-divergence minimization framework.  ...  We unfortunately find that f-divergence minimization through reinforcement learning is susceptible to numerical instabilities.  ...  ., and Levine, S. A connection between generative adversarial networks, inverse reinforcement learning, and energy-based models. arXiv preprint arXiv:1611.03852, 2016a.  ... 
arXiv:2006.10810v1 fatcat:zobyenu2rbfqfbjlk22yhus74u

Longitudinal Face Aging in the Wild - Recent Deep Learning Approaches [article]

Chi Nhan Duong, Khoa Luu, Kha Gia Quach, Tien D. Bui
2018 arXiv   pre-print
In this paper, we aim to give a review of recent developments of modern deep learning based approaches, i.e. Deep Generative Models, for Face Aging task.  ...  Face Aging has raised considerable attentions and interest from the computer vision community in recent years.  ...  DEEP LEARNING (DL), LOG-LIKELIHOOD (LL), INVERSE REINFORCEMENT LEARNING (IRL), PROBABILISTIC GRAPHICAL MODELS (PGM), ADVERSARIAL (ADV) Method Approach Architecture Loss Function Non -Linearity  ... 
arXiv:1802.08726v1 fatcat:vomix3nu6nfsdmnbfhv6ivjjeu

Learning to Prevent Leakage: Privacy-Preserving Inference in the Mobile Cloud [article]

Shuang Zhang, Liyao Xiang, Congcong Li, Yixuan Wang, Quanshi Zhang, Wei Wang, Bo Li
2021 arXiv   pre-print
As deep learning tasks are mostly computation-intensive, it has become a trend to process raw data on devices and send the deep neural network (DNN) features to the cloud, where the features are further  ...  The framework aims to learn a policy to modify the base DNNs to prevent information leakage while maintaining high inference accuracy.  ...  To achieve a better tradeoff among the objectives of privacy, accuracy and resource utilization, we present a reinforcement learning based optimizer searching for the optimal neural network transformation  ... 
arXiv:1912.08421v2 fatcat:um3zb6cbi5cjxmpueluw556v24

Reinforcement Learning for IoT Security: A Comprehensive Survey

Aashma Uprety, Danda B. Rawat
2020 IEEE Internet of Things Journal  
In this paper, we present an comprehensive survey of different types of cyber-attacks against different IoT systems and then we present reinforcement learning and deep reinforcement learning based security  ...  With this paper, readers can have a more thorough understanding of IoT security attacks and countermeasures using Reinforcement Learning, as well as research trends in this area.  ...  REINFORCEMENT LEARNING IN CYBER PHYSICAL SYSTEMS A. Security in Smart Grid Smart Grid is an intelligent system to generate and distribute energy in a distributed manner.  ... 
doi:10.1109/jiot.2020.3040957 fatcat:qtm2emhqmjhlncyczuij6nw3oa

Learning to infer in recurrent biological networks [article]

Ari S. Benjamin, Konrad P. Kording
2021 arXiv   pre-print
A popular theory of perceptual processing holds that the brain learns both a generative model of the world and a paired recognition model using variational Bayesian inference.  ...  We illustrate the idea on recurrent neural networks trained to model image and video datasets.  ...  Related work in generative modeling Several generative modeling papers have approached representation learning by adversarially matching joint distributions of latent vectors and inputs.  ... 
arXiv:2006.10811v2 fatcat:jlatutwa6vb3znm4ms4mafyrky

From internal models toward metacognitive AI [article]

Mitsuo Kawato
2021 arXiv   pre-print
The model comprises a modular hierarchical reinforcement-learning architecture of parallel and layered, generative-inverse model pairs.  ...  Based on mismatches between computations by generative and inverse models, as well as reward prediction errors, CRMN computes a "responsibility signal" that gates selection and learning of pairs in perception  ...  The model comprises a modular hierarchical reinforcement-learning architecture of parallel and layered, generative-inverse model pairs.  ... 
arXiv:2109.12798v2 fatcat:7foykaimirblrcxg3anqjo2jfm

A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications [article]

Jie Gui, Zhenan Sun, Yonggang Wen, Dacheng Tao, Jieping Ye
2020 arXiv   pre-print
Generative adversarial networks (GANs) are a hot research topic recently. GANs have been widely studied since 2014, and a large number of algorithms have been proposed.  ...  Furthermore, GANs have been combined with other machine learning algorithms for specific applications, such as semi-supervised learning, transfer learning, and reinforcement learning.  ...  The authors also would like to thank the helpful discussions with group members of Umich Yelab and Foreseer research group.  ... 
arXiv:2001.06937v1 fatcat:4iqb2vnhezgjnphfv3taej7vbu

UAV Autonomous Aerial Combat Maneuver Strategy Generation with Observation Error Based on State-Adversarial Deep Deterministic Policy Gradient and Inverse Reinforcement Learning

Weiren Kong, Deyun Zhou, Zhen Yang, Yiyang Zhao, Kai Zhang
2020 Electronics  
At the same time, a reward shaping method based on maximum entropy (MaxEnt) inverse reinforcement learning algorithm (IRL) is proposed to improve the aerial combat strategy generation algorithm's efficiency  ...  Meanwhile, we propose a novel autonomous aerial combat maneuver strategy generation algorithm with high-performance and high-robustness based on state-adversarial deep deterministic policy gradient algorithm  ...  Reward Shaping Using Inverse Reinforcement Learning Reward Shaping Due to the learning efficiency of deep reinforcement learning, reinforcement learning has been restricting its practical application  ... 
doi:10.3390/electronics9071121 fatcat:cbu5qoteuzdovkyhhyyo2q6db4
« Previous Showing results 1 — 15 out of 4,188 results