Filters








9,481 Hits in 2.6 sec

Hash Layers For Large Sparse Models [article]

Stephen Roller, Sainbayar Sukhbaatar, Arthur Szlam, Jason Weston
<span title="2021-07-20">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We investigate the training of sparse layers that use different parameters for different inputs based on hashing in large Transformer models.  ...  We show our approach works well both on large language modeling and dialogue tasks, and on downstream fine-tuning tasks.  ...  Conclusion We have introduced a simple and efficient approach to sparse models in the Transformers-for-NLP setting based on hash layers.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2106.04426v3">arXiv:2106.04426v3</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/do7aoi2flja4dpfgmrguygxypu">fatcat:do7aoi2flja4dpfgmrguygxypu</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210722120308/https://arxiv.org/pdf/2106.04426v3.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/06/11/0611d2f2ea6a3c8fb8534f42758a5a3e9c7bc8fe.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2106.04426v3" 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>

Model Size Reduction Using Frequency Based Double Hashing for Recommender Systems [article]

Caojin Zhang, Yicun Liu, Yuanpu Xie, Sofia Ira Ktena, Alykhan Tejani, Akshay Gupta, Pranay Kumar Myana, Deepak Dilipkumar, Suvadip Paul, Ikuhiro Ihara, Prasang Upadhyaya, Ferenc Huszar (+1 others)
<span title="2020-07-28">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In this paper, we propose a hybrid hashing method to combine frequency hashing and double hashing techniques for model size reduction, without compromising performance.  ...  The large model size usually entails a cost, in the range of millions of dollars, for storage and communication with the inference services.  ...  By reducing the size of embedding layers, we can reduce the large memory capacity challenges in production-scale models.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2007.14523v1">arXiv:2007.14523v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/oh7xmkcu5jdgdk6uacpo2aiyle">fatcat:oh7xmkcu5jdgdk6uacpo2aiyle</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200806110852/https://arxiv.org/pdf/2007.14523v1.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] </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2007.14523v1" 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>

Scalable and Sustainable Deep Learning via Randomized Hashing [article]

Ryan Spring, Anshumali Shrivastava
<span title="2016-12-05">2016</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
A unique property of the proposed hashing based back-propagation is that the updates are always sparse.  ...  Our new algorithm for deep learning reduces the overall computational cost of forward and back-propagation by operating on significantly fewer (sparse) nodes.  ...  l -Layer l Hash Function // HT l -Layer l Hash Tables // AS l -Layer l Active Set // θ l AS ∈ W l AS , b l AS -Layer l Active Set parameters Randomly initialize parameters W l , b l for each layer l HF  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1602.08194v2">arXiv:1602.08194v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/fo2pjpzsivgzxnmxurmlalfvda">fatcat:fo2pjpzsivgzxnmxurmlalfvda</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191021132009/https://arxiv.org/pdf/1602.08194v2.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/b9/46/b9464073dd3073cd464c9c93513690ecc581fff6.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1602.08194v2" 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>

Understanding Training Efficiency of Deep Learning Recommendation Models at Scale [article]

Bilge Acun, Matthew Murphy, Xiaodong Wang, Jade Nie, Carole-Jean Wu, Kim Hazelwood
<span title="2020-11-11">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
GPU performance and efficiency of these recommendation models are largely affected by model architecture configurations such as dense and sparse features, MLP dimensions.  ...  The use of GPUs has proliferated for machine learning workflows and is now considered mainstream for many deep learning models.  ...  ACKNOWLEDGEMENTS We would like to thank Facebook colleagues, especially Shunting Zhang, Hassan Eslami, Chenguang Xi, Manoj Krishnan, Jiyan Yang for the discussions and feedback on this work.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2011.05497v1">arXiv:2011.05497v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/6nddgrsi25fbhiwxswnz263pda">fatcat:6nddgrsi25fbhiwxswnz263pda</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20201113032619/https://arxiv.org/pdf/2011.05497v1.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/c4/a2c4647ce555fa71a7e39f94c08a1329bbd0da85.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2011.05497v1" 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>

Cluster-Former: Clustering-based Sparse Transformer for Long-Range Dependency Encoding [article]

Shuohang Wang, Luowei Zhou, Zhe Gan, Yen-Chun Chen, Yuwei Fang, Siqi Sun, Yu Cheng, Jingjing Liu
<span title="2021-06-07">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
The proposed framework is pivoted on two unique types of Transformer layer: Sliding-Window Layer and Cluster-Former Layer, which encode local sequence information and global context jointly and iteratively  ...  In this paper, we propose Cluster-Former, a novel clustering-based sparse Transformer to perform attention across chunked sequences.  ...  Question Answering: We initialize our models with RoBERTa-large that has 24 Transformer layers, 16 heads per layer and hidden state dimension of 1024.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2009.06097v2">arXiv:2009.06097v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/stsmr7vwone2jfm3fo7t2t2dku">fatcat:stsmr7vwone2jfm3fo7t2t2dku</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200919190543/https://arxiv.org/pdf/2009.06097v1.pdf" title="fulltext PDF download [not primary version]" 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] <span style="color: #f43e3e;">&#10033;</span> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2009.06097v2" 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>

Efficient Language Modeling with Sparse all-MLP [article]

Ping Yu, Mikel Artetxe, Myle Ott, Sam Shleifer, Hongyu Gong, Ves Stoyanov, Xian Li
<span title="2022-03-16">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
and HASH Layers) as well as dense Transformers and all-MLPs.  ...  The proposed sparse all-MLP improves language modeling perplexity and obtains up to 2× improvement in training efficiency compared to both Transformer-based MoEs (GShard, Switch Transformer, Base Layers  ...  Although Base Layers and HASH Layers perform well when trained on small models and datasets, we find stability issues on large-scale training.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2203.06850v2">arXiv:2203.06850v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/vynobpauhfg5dkpytfrkgk77na">fatcat:vynobpauhfg5dkpytfrkgk77na</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220320093909/https://arxiv.org/pdf/2203.06850v2.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/52/cc/52ccaa67a8d0d9025bba8cfb1da1e5650527c0d3.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2203.06850v2" 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>

Learning a Neural Diff for Speech Models [article]

Jonathan Macoskey, Grant P. Strimel, Ariya Rastrow
<span title="2021-08-17">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We present neural update approaches for release of subsequent speech model generations abiding by a data budget.  ...  In this work we address one of these constraints, namely over-the-network data budgets for transferring models from server to device.  ...  Sparse Hash Sparse Hash Static CR 1 10 10 1 20 40 20 40 inf hours BL BLC Diff BLC Diff # dom.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2108.01561v2">arXiv:2108.01561v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/vtngciz3ivalpovpdoefw5ll6m">fatcat:vtngciz3ivalpovpdoefw5ll6m</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210829080734/https://arxiv.org/pdf/2108.01561v2.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/0c/55/0c5533fecd7b069fd8e231ef7c47c52a53fc01a7.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2108.01561v2" 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 Semantic-Preserving Deep Hashing Model for Multi-Label Remote Sensing Image Retrieval

Qimin Cheng, Haiyan Huang, Lan Ye, Peng Fu, Deqiao Gan, Yuzhuo Zhou
<span title="">2021</span> <i title="MDPI AG"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/kay2tsbijbawliu45dnhvyvgsq" style="color: black;">Remote Sensing</a> </i> &nbsp;
we propose a new semantic-preserving deep hashing model for multi-label remote sensing image retrieval.  ...  Our model consists of three main components: (1) a convolutional neural network to extract image features; (2) a hash layer to generate binary codes; (3) a new loss function to better maintain the multi-label  ...  Acknowledgments: The authors would like to thank reviewers for reviewing this paper and providing important feedback throughout its development.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.3390/rs13244965">doi:10.3390/rs13244965</a> <a target="_blank" rel="external noopener" href="https://doaj.org/article/48dbc4fad3f2499c98690893c1e4f032">doaj:48dbc4fad3f2499c98690893c1e4f032</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/tazkazpeibewhpwmfmuhwea5wy">fatcat:tazkazpeibewhpwmfmuhwea5wy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220429124634/https://mdpi-res.com/d_attachment/remotesensing/remotesensing-13-04965/article_deploy/remotesensing-13-04965.pdf?version=1638870905" 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/13/8f13fcf0d003fc2f21de8b00b46e185ccf587702.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.3390/rs13244965"> <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>

Distributed Hierarchical GPU Parameter Server for Massive Scale Deep Learning Ads Systems [article]

Weijie Zhao, Deping Xie, Ronglai Jia, Yulei Qian, Ruiquan Ding, Mingming Sun, Ping Li
<span title="2020-03-12">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
For example, a sponsored online advertising system can contain more than 10^11 sparse features, making the neural network a massive model with around 10 TB parameters.  ...  We propose a hierarchical workflow that utilizes GPU High-Bandwidth Memory, CPU main memory and SSD as 3-layer hierarchical storage.  ...  For the parameters in the embedding layer, only a subset of them is used and will be updated for this sparse input (we call them "sparse parameters" for convenience).  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2003.05622v1">arXiv:2003.05622v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/kfl2uv7oarfsfa7zpkgps76h6e">fatcat:kfl2uv7oarfsfa7zpkgps76h6e</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200320182729/https://arxiv.org/pdf/2003.05622v1.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] </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2003.05622v1" 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>

Sketching and Neural Networks [article]

Amit Daniely and Nevena Lazic and Yoram Singer and Kunal Talwar
<span title="2016-04-19">2016</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
High-dimensional sparse data present computational and statistical challenges for supervised learning.  ...  We propose compact linear sketches for reducing the dimensionality of the input, followed by a single layer neural network.  ...  Acknowledgements We would like to thank Amir Globerson for numerous fruitful discussion and help with an early version of the manuscript.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1604.05753v1">arXiv:1604.05753v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/v2tl7w7zn5csvht5jqrus2bdsa">fatcat:v2tl7w7zn5csvht5jqrus2bdsa</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200828070135/https://arxiv.org/pdf/1604.05753v1.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/18/09/18099f0d3b856a301c3d4bd47024f91ce17fec93.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1604.05753v1" 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>

Survey Paper on Generating Correlation among Different Modalities by Using Parallel Processing for Cross-Media Retrieval

Rokkam SrikanthReddy
<span title="2017-08-30">2017</span> <i title="International Journal for Research in Applied Science and Engineering Technology (IJRASET)"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/hsp44774azcezeyiq4kuzpfh5a" style="color: black;">International Journal for Research in Applied Science and Engineering Technology</a> </i> &nbsp;
Hashing methods are useful for performing variety of tasks in recent years. Various hashing approaches have been performing retrieve the cross-media information.  ...  To retrieving cross-media information from large data sets it causes difficulties to process and retrieve the information at run time.  ...  Latent Semantic Sparse Hashing for Cross-Modal Similarity Search G. Ding, J. Zhou and Y.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.22214/ijraset.2017.8309">doi:10.22214/ijraset.2017.8309</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/5j65haeztfdedoljvqrczo2pxq">fatcat:5j65haeztfdedoljvqrczo2pxq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20180603022019/http://ijraset.com/fileserve.php?FID=9732" 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/e2/c3/e2c385221c93034c68e7855575fcc8135fca1202.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.22214/ijraset.2017.8309"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="unlock alternate icon" style="background-color: #fb971f;"></i> Publisher / doi.org </button> </a>

SLIDE : In Defense of Smart Algorithms over Hardware Acceleration for Large-Scale Deep Learning Systems [article]

Beidi Chen, Tharun Medini, James Farwell, Sameh Gobriel, Charlie Tai, Anshumali Shrivastava
<span title="2020-03-01">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
To get around the costly computations associated with large models and data, the community is increasingly investing in specialized hardware for model training.  ...  We provide codes and scripts for reproducibility.  ...  ACKNOWLEDGEMENTS The work was supported by NSF-1652131, nsf-bigdata 1838177, AFOSR-YIPFA9550-18-1-0152, Amazon Research Award, and ONR BRC grant for Randomized Numerical Linear Algebra.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1903.03129v2">arXiv:1903.03129v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/pbeouobwabehdmefvveage75vu">fatcat:pbeouobwabehdmefvveage75vu</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200321201535/https://arxiv.org/pdf/1903.03129v2.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/c4/7b/c47b89a2b57950bb273a7992592631e5524fec61.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1903.03129v2" 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>

Accelerating SLIDE Deep Learning on Modern CPUs: Vectorization, Quantizations, Memory Optimizations, and More [article]

Shabnam Daghaghi, Nicholas Meisburger, Mengnan Zhao, Yong Wu, Sameh Gobriel, Charlie Tai, Anshumali Shrivastava
<span title="2021-03-06">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
SLIDE is a C++ implementation of a sparse hash table based back-propagation, which was shown to be significantly faster than GPUs in training hundreds of million parameter neural models.  ...  Our experiments are focused on large (hundreds of millions of parameters) recommendation and NLP models.  ...  We apply hash functions for the last layer where we have the computational bottleneck.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2103.10891v1">arXiv:2103.10891v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/kvi4fszq4vampgsztwo52omc34">fatcat:kvi4fszq4vampgsztwo52omc34</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210331070813/https://arxiv.org/pdf/2103.10891v1.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/28/da28165f30e1e8ad91521c2a99dc59681176162f.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2103.10891v1" 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>

Learning A Deep ℓ_∞ Encoder for Hashing [article]

Zhangyang Wang, Yingzhen Yang, Shiyu Chang, Qing Ling, Thomas S. Huang
<span title="2016-04-06">2016</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We then investigate the effective use of the proposed model in the application of hashing, by coupling the proposed encoders under a supervised pairwise loss, to develop a Deep Siamese ℓ_∞ Network, which  ...  Such a structural prior acts as an effective network regularization, and facilitates the model initialization.  ...  We re-implement the encoder parts of NNH and SNNH, with three hidden layers (i.e, two unfolded stages for LISTA), so that all three deep hashing models have the same depth 3 .  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1604.01475v1">arXiv:1604.01475v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/isspjfxi6bf3lkrknx6nfrxzqq">fatcat:isspjfxi6bf3lkrknx6nfrxzqq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200910172623/https://arxiv.org/pdf/1604.01475v1.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/bc/41/bc41e60667bcef0f9c0ef3b19e957924b125d5a0.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1604.01475v1" 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>

Compressing Neural Networks with the Hashing Trick [article]

Wenlin Chen and James T. Wilson and Stephen Tyree and Kilian Q. Weinberger and Yixin Chen
<span title="2015-04-19">2015</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
; however mobile devices are designed with very little memory and cannot store such large models.  ...  As deep nets are increasingly used in applications suited for mobile devices, a fundamental dilemma becomes apparent: the trend in deep learning is to grow models to absorb ever-increasing data set sizes  ...  Feature Hashing Learning under memory constraints has previously been explored in the context of large-scale learning for sparse data sets.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1504.04788v1">arXiv:1504.04788v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/lpwthabrt5ffjle2jcv3652gjq">fatcat:lpwthabrt5ffjle2jcv3652gjq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191017063342/https://arxiv.org/pdf/1504.04788v1.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/b9/6f/b96fb3bba20047f73a908c040cbab6e4fafe3beb.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1504.04788v1" 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>
&laquo; Previous Showing results 1 &mdash; 15 out of 9,481 results