Filters








1,749 Hits in 4.1 sec

IMPACT: Importance Weighted Asynchronous Architectures with Clipped Target Networks [article]

Michael Luo, Jiahao Yao, Richard Liaw, Eric Liang, Ion Stoica
<span title="2020-01-23">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
IMPACT extends IMPALA with three changes: a target network for stabilizing the surrogate objective, a circular buffer, and truncated importance sampling.  ...  To accelerate training, practitioners often turn to distributed reinforcement learning architectures to parallelize and accelerate the training process.  ...  The new agent, Importance Weighted Asynchronous Architectures with Clipped Target Networks (IMPACT), mitigates this inherent mismatch.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1912.00167v3">arXiv:1912.00167v3</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/onamkixxnnaktfaz2u3vx4qkli">fatcat:onamkixxnnaktfaz2u3vx4qkli</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200830104837/https://arxiv.org/pdf/1912.00167v3.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/a2/e8/a2e8e374b0d948effa3fa626f3cc3f489c18c1d5.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1912.00167v3" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Application of Deep Reinforcement Learning in Traffic Signal Control: An Overview and Impact of Open Traffic Data

Martin Gregurić, Miroslav Vujić, Charalampos Alexopoulos, Mladen Miletić
<span title="2020-06-10">2020</span> <i title="MDPI AG"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/smrngspzhzce7dy6ofycrfxbim" style="color: black;">Applied Sciences</a> </i> &nbsp;
Best practices are provided for choosing the adequate DRL model, hyper-parameters tuning, and model architecture design.  ...  Finally, this paper provides a discussion about the importance of the open traffic data concept for the extensive application of DRL in the real world ATSC.  ...  Keeping the fixed weights in target DNN model for a predefined period of time ensures a temporally static Q-value target.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.3390/app10114011">doi:10.3390/app10114011</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/vnhrunfhpbgtrmky7pcvdg223a">fatcat:vnhrunfhpbgtrmky7pcvdg223a</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200611045837/https://res.mdpi.com/d_attachment/applsci/applsci-10-04011/article_deploy/applsci-10-04011.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/4c/60/4c6054773eca76f1cc1622417711e4e545a12e77.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.3390/app10114011"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="unlock alternate icon" style="background-color: #fb971f;"></i> mdpi.com </button> </a>

Trajectory Based Prioritized Double Experience Buffer for Sample-Efficient Policy Optimization

Shengxiang Li, Ou Li, Guangyi Liu, Siyuan Ding, Yijie Bai
<span title="">2021</span> <i title="Institute of Electrical and Electronics Engineers (IEEE)"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/q7qi7j4ckfac7ehf3mjbso4hne" style="color: black;">IEEE Access</a> </i> &nbsp;
This paper introduces a novel policy gradient method to improve the sample efficiency via a pair of trajectory based prioritized replay buffers and reduce the variance in training with a target network  ...  whose weights are updated in a "soft" manner.  ...  To prevent this, we pulls back larger importance sampling ratio via clipping the importance sampling with target network.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/access.2021.3097357">doi:10.1109/access.2021.3097357</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/evq3kgrsxnfm3cjnvwpbutw7mu">fatcat:evq3kgrsxnfm3cjnvwpbutw7mu</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210804072609/https://ieeexplore.ieee.org/ielx7/6287639/9312710/09486881.pdf?tp=&amp;arnumber=9486881&amp;isnumber=9312710&amp;ref=" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/bf/d1/bfd19d603ac11954bacafd948055c2b43ff9a14b.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/access.2021.3097357"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="unlock alternate icon" style="background-color: #fb971f;"></i> ieee.com </button> </a>

Autoregressive Convolutional Neural Networks for Asynchronous Time Series [article]

Mikołaj Bińkowski, Gautier Marti, Philippe Donnat
<span title="2018-06-12">2018</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We propose Significance-Offset Convolutional Neural Network, a deep convolutional network architecture for regression of multivariate asynchronous time series.  ...  It involves an AR-like weighting system, where the final predictor is obtained as a weighted sum of adjusted regressors, while the weights are datadependent functions learnt through a convolutional network  ...  This means that the significance network has crucial impact on the performance, which is in-line with the potential drawbacks of the LSTM network discussed in Section 3: obtaining proper weights for the  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1703.04122v4">arXiv:1703.04122v4</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/aqiz75ukm5drxkqspiyq3fhf2a">fatcat:aqiz75ukm5drxkqspiyq3fhf2a</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200929013754/https://arxiv.org/pdf/1703.04122v4.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/90/ac/90ac1a43437ae614b4eb587abcf46738c783dfc1.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1703.04122v4" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

neXtream: A Multi-Device, Social Approach to Video Content Consumption

ReeD Martin, Ana Luisa Santos, Mike Shafran, Henry Holtzman, Marie-Jose Montpetit
<span title="">2010</span> <i title="IEEE"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/iziq5tdnabbczezynz7xdk7uma" style="color: black;">2010 7th IEEE Consumer Communications and Networking Conference</a> </i> &nbsp;
At the same time, the varying methods of viewing, interacting with, and sharing content have diverged.  ...  The paper presents the system concept, theory, and architecture, and describes the developed prototype.  ...  This tab also provides an indicator of communication with friends around the current video clip, allowing the user to comment asynchronously with friends about the currently playing video clip.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/ccnc.2010.5421599">doi:10.1109/ccnc.2010.5421599</a> <a target="_blank" rel="external noopener" href="https://dblp.org/rec/conf/ccnc/MartinSSHM10.html">dblp:conf/ccnc/MartinSSHM10</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/53o3l7pzrbcgtjh7tjnyuypgnm">fatcat:53o3l7pzrbcgtjh7tjnyuypgnm</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220114094921/https://dspace.mit.edu/bitstream/handle/1721.1/62200/Martin-2010-neXtream%20a%20multi-device,%20social%20approach%20to%20video%20content%20consumption.pdf;jsessionid=D67DB9D35F155454A8CB9D97D2CAF7E4?sequence=2" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/d6/f9/d6f9844d474a04c9a2def64561001cc7c71ea3f0.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/ccnc.2010.5421599"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> ieee.com </button> </a>

Federated Action Recognition on Heterogeneous Embedded Devices [article]

Pranjal Jain, Shreyas Goenka, Saurabh Bagchi, Biplab Banerjee, Somali Chaterji
<span title="2021-07-18">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We empirically show on a testbed of heterogeneous embedded devices that we can perform action recognition with comparable accuracy to the two baselines above, while our asynchronous learning strategy reduces  ...  In this work, we enable clients with limited computing power to perform action recognition, a computationally heavy task.  ...  As can be seen in Table III , the asynchronous training achieves higher accuracy for both per-clip and per-video metrics. This emphasizes the importance of our design of dealing C.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2107.12147v1">arXiv:2107.12147v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/o6y4p3w3drfhpjebtb5nwjmana">fatcat:o6y4p3w3drfhpjebtb5nwjmana</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210730000443/https://arxiv.org/pdf/2107.12147v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/d1/50/d150e40d16c7ece4038e4a861ee1a896e53aede6.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2107.12147v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Physiology and Impact of Horizontal Connections in Rat Neocortex

Philipp Schnepel, Arvind Kumar, Mihael Zohar, Ad Aertsen, Clemens Boucsein
<span title="2014-11-19">2014</span> <i title="Oxford University Press (OUP)"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/eeg67t2wzfd3dpiicxsttg3cxi" style="color: black;">Cerebral Cortex</a> </i> &nbsp;
However, recent studies suggest that the bulk of axons targeting pyramidal neurons most likely originate from outside this local range, emphasizing the importance of horizontal connections.  ...  Implementing our data into a spiking neuronal network model shows that more horizontal connections promote robust asynchronous ongoing activity states and reduce noise correlations in stimulus-induced  ...  the impact of horizontal connections on key parameters of network activity.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1093/cercor/bhu265">doi:10.1093/cercor/bhu265</a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pubmed/25410428">pmid:25410428</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/6kcb7qm7qzahzpvppekvc7eelm">fatcat:6kcb7qm7qzahzpvppekvc7eelm</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20190308024705/http://pdfs.semanticscholar.org/eb7a/58aad3f4a66774e380cc1fcf809bc32b781a.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/eb/7a/eb7a58aad3f4a66774e380cc1fcf809bc32b781a.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1093/cercor/bhu265"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> oup.com </button> </a>

AFLPC: An Asynchronous Federated Learning Privacy-Preserving Computing Model Applied to 5G-V2X

Jie Huang, Cheng Xu, Zhaohua Ji, Shan Xiao, Teng Liu, Nan Ma, Qinghui Zhou, Muhammad Arif
<span title="2022-03-08">2022</span> <i title="Hindawi Limited"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/sdme5pnua5auzcsjgqmqefb66m" style="color: black;">Security and Communication Networks</a> </i> &nbsp;
A weight-based asynchronous federated learning aggregation update method is proposed to reasonably control the proportion of parameters submitted by users with different training speeds in the aggregation  ...  The advantages of low delay of 5G network should be better utilized in the vehicle-road cooperative system.  ...  Figure 1 : 1 Figure 1: Asynchronous Federated Learning System Architecture, in which multiple terminals containing different user populations are connected to each other over 5G networks, and each terminal  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1155/2022/9334943">doi:10.1155/2022/9334943</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/nqw4mo5tifgmrfpguosr3f6rea">fatcat:nqw4mo5tifgmrfpguosr3f6rea</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220313200851/https://downloads.hindawi.com/journals/scn/2022/9334943.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/18/64/18642da9d5707865efd24d99e258af805680649f.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1155/2022/9334943"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="unlock alternate icon" style="background-color: #fb971f;"></i> hindawi.com </button> </a>

Zero-touch Continuous Network Slicing Control via Scalable Actor-Critic Learning [article]

Farhad Rezazadeh, Hatim Chergui, Christos Verikoukis
<span title="2021-01-17">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
The paper defines and corroborates via extensive experimental results a zero-touch network slicing scheme with a multi-objective approach where the central server learns continuously to accumulate the  ...  Moreover, we pursue a state-action return distribution learning approach with the proposed replay policy and reward-penalty mechanisms.  ...  The prioritized actor-learner architecture optimized for the network slicing environment. asynchronous actor-learner optimized experience replay architecture for the network slicing environment.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2101.06654v1">arXiv:2101.06654v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/ijtxmfamifdvno2vjkjcqb2m5e">fatcat:ijtxmfamifdvno2vjkjcqb2m5e</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210120105401/https://arxiv.org/pdf/2101.06654v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/da/99/da997b1534940fdf31e3a2a1fad629104fb6407e.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2101.06654v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

MQBench: Towards Reproducible and Deployable Model Quantization Benchmark [article]

Yuhang Li, Mingzhu Shen, Jian Ma, Yan Ren, Mingxin Zhao, Qi Zhang, Ruihao Gong, Fengwei Yu, Junjie Yan
<span title="2022-01-25">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
However, in QAT folding BN into weights with asynchronous statistics will produce different quantized weights, further magnifying the training instability.  ...  If the target hardware or the network architectures are not met before, we recommend using LSQ since it has the best average performance in history (Fig. 6 ).  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2111.03759v2">arXiv:2111.03759v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/k5woc6az6zfgpcgx5dxgzhkfou">fatcat:k5woc6az6zfgpcgx5dxgzhkfou</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220128084316/https://arxiv.org/pdf/2111.03759v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/58/fb/58fb1f1459e50eaa969da031d9ebd7d4e4747898.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2111.03759v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures [article]

Lasse Espeholt, Hubert Soyer, Remi Munos, Karen Simonyan, Volodymir Mnih, Tom Ward, Yotam Doron, Vlad Firoiu, Tim Harley, Iain Dunning, Shane Legg, Koray Kavukcuoglu
<span title="2018-06-28">2018</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We have developed a new distributed agent IMPALA (Importance Weighted Actor-Learner Architecture) that not only uses resources more efficiently in single-machine training but also scales to thousands of  ...  In this work we aim to solve a large collection of tasks using a single reinforcement learning agent with a single set of parameters.  ...  We propose the Importance Weighted Actor-Learner Architecture (IMPALA) shown in Figure 1 .  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1802.01561v3">arXiv:1802.01561v3</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/q4io6ns52vgejholewhzdshb2y">fatcat:q4io6ns52vgejholewhzdshb2y</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191025110911/https://arxiv.org/pdf/1802.01561v3.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/8c/c0/8cc09844e96f48e23bdd3bd141505962cea3f46c.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1802.01561v3" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Don't Judge Me by My Face : An Indirect Adversarial Approach to Remove Sensitive Information From Multimodal Neural Representation in Asynchronous Job Video Interviews [article]

Léo Hemamou, Arthur Guillon, Jean-Claude Martin, Chloé Clavel
<span title="2021-10-18">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Comparing our approach to a standard baseline on a public dataset with gender and ethnicity annotations, we show that it effectively removes sensitive information from the main network.  ...  Recently, adversarial methods have been proved to effectively remove sensitive information from the latent representation of neural networks.  ...  Base network for hireability. Our approach was not designed with a specific architecture in mind and could be used on any deep learning algorithm trained for hireability.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2110.09424v1">arXiv:2110.09424v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/iknfspsudbdvjdqfwvpvvcg5t4">fatcat:iknfspsudbdvjdqfwvpvvcg5t4</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20211021205042/https://arxiv.org/pdf/2110.09424v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/19/ce/19ce3ad3227f3d9dbd3a5ee990f58e7417bb2b2e.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2110.09424v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Pre-training with Non-expert Human Demonstration for Deep Reinforcement Learning [article]

Gabriel V. de la Cruz, Yunshu Du, Matthew E. Taylor
<span title="2018-12-21">2018</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Deep reinforcement learning (deep RL) has achieved superior performance in complex sequential tasks by using deep neural networks as function approximators to learn directly from raw input images.  ...  We leverage supervised learning to pre-train on a small set of non-expert human demonstrations and empirically evaluate our approach using the asynchronous advantage actor-critic algorithms (A3C) in the  ...  The use of a target network, an experience replay memory, and the reward clipping are essential to stabilizing learning.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1812.08904v1">arXiv:1812.08904v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/khfqtotuqbglpab2yy4x3uhusy">fatcat:khfqtotuqbglpab2yy4x3uhusy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200826162153/https://arxiv.org/pdf/1812.08904v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/ac/1a/ac1ad3a421c5db0d1f29b815359312a22357d40b.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1812.08904v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

GOAT: GPU Outsourcing of Deep Learning Training With Asynchronous Probabilistic Integrity Verification Inside Trusted Execution Environment [article]

Aref Asvadishirehjini
<span title="2020-10-17">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Machine learning models based on Deep Neural Networks (DNNs) are increasingly deployed in a wide range of applications ranging from self-driving cars to COVID-19 treatment discovery.  ...  Yet, no existing approach scales up to support realistic integrity-preserving DNN model training for heavy workloads (deep architectures and millions of training examples) without sustaining a significant  ...  F.3 Impact of Gradient Clipping for Honest Trainers One important question is whether the gradient clipping used to prevent attacker to change parameters in a given mini-batch update would have performance  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2010.08855v1">arXiv:2010.08855v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/hu5dlsb6erhrfjkb2ql7worg74">fatcat:hu5dlsb6erhrfjkb2ql7worg74</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20201024180138/https://arxiv.org/pdf/2010.08855v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/43/18/431848801e55bd82ad5a38670d253cea32c624ca.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2010.08855v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

A Survey of Multi-Task Deep Reinforcement Learning

Nelson Vithayathil Varghese, Qusay H. Mahmoud
<span title="2020-08-22">2020</span> <i title="MDPI AG"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/ikdpfme5h5egvnwtvvtjrnntyy" style="color: black;">Electronics</a> </i> &nbsp;
recent solutions, namely DISTRAL (DIStill & TRAnsfer Learning), IMPALA(Importance Weighted Actor-Learner Architecture) and PopArt that aim to address core challenges such as scalability, distraction dilemma  ...  Undoubtedly, the inception of deep reinforcement learning has played a vital role in optimizing the performance of reinforcement learning-based intelligent agents with model-free based approaches.  ...  The PopArt model was designed based on the original IMPALA (importance weighted actor-learner architecture) architecture model with the combination of multiple convolutional neural network layers with  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.3390/electronics9091363">doi:10.3390/electronics9091363</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/cohk2pukzbgbfizarqweuw45oe">fatcat:cohk2pukzbgbfizarqweuw45oe</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200904132532/https://res.mdpi.com/d_attachment/electronics/electronics-09-01363/article_deploy/electronics-09-01363.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/8f/7f/8f7fe421f4db34d8a3627894577af3e0ac4ee51f.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.3390/electronics9091363"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="unlock alternate icon" style="background-color: #fb971f;"></i> mdpi.com </button> </a>
&laquo; Previous Showing results 1 &mdash; 15 out of 1,749 results