Filters








104,450 Hits in 5.2 sec

Synthetic Examples Improve Generalization for Rare Classes [article]

Sara Beery, Yang Liu, Dan Morris, Jim Piavis, Ashish Kapoor, Markus Meister, Neel Joshi, Pietro Perona
<span title="2019-05-14">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Our experiments reveal that synthetic data can considerably reduce error rates for classes that are rare, that as the amount of simulated data is increased, accuracy on the target class improves, and that  ...  The ability to detect and classify rare occurrences in images has important applications - for example, counting rare and endangered species when studying biodiversity, or detecting infrequent traffic  ...  Compute provided by Microsoft AI for Earth and AWS.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1904.05916v2">arXiv:1904.05916v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/qielv4aa7fcxvpwacvgnkoy43m">fatcat:qielv4aa7fcxvpwacvgnkoy43m</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200928031811/https://arxiv.org/pdf/1904.05916v2.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/b6/dc/b6dcfc2e4d51112c6093a84d5abfd90260ff6e93.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1904.05916v2" 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>

Road images augmentation with synthetic traffic signs using neural networks

A.S. Konushin, B.V. Faizov, V.I. Shakhuro
<span title="">2021</span> <i title="Samara State National Research University"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/b3lfu4665ndmbj6r3a7h4u4sim" style="color: black;">Computer Optics</a> </i> &nbsp;
We propose three methods for making synthetic signs consistent with a scene in appearance. These methods are based on modern generative adversarial network (GAN) architectures.  ...  However, the state of the art methods works only for frequent sign classes, which are well represented in training datasets. We consider the task of rare traffic sign detection and classification.  ...  It showed that our method for generating synthetic training samples improved quality for detector and classifier of traffic signs. For all sign classes, recognition quality improved.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.18287/2412-6179-co-859">doi:10.18287/2412-6179-co-859</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/grncol6wzvehlb5rcvgxfakery">fatcat:grncol6wzvehlb5rcvgxfakery</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220315180100/http://www.computeroptics.ru/KO/PDF/KO45-5/450513.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/9a/e0/9ae0b0115323eceecd59505775d1b17acbbf6d67.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.18287/2412-6179-co-859"> <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>

Road images augmentation with synthetic traffic signs using neural networks [article]

Anton Konushin, Boris Faizov, Vlad Shakhuro
<span title="2021-01-13">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We propose three methods for making synthetic signs consistent with a scene in appearance. These methods are based on modern generative adversarial network (GAN) architectures.  ...  However, the state of the art methods works only for frequent sign classes, which are well represented in training datasets. We consider the task of rare traffic sign detection and classification.  ...  This way of generating synthetic data allows increase training set with new examples of rare classes with the correct geometric position.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2101.04927v1">arXiv:2101.04927v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/v5qvswlskrfldeyhxnqzdjvpem">fatcat:v5qvswlskrfldeyhxnqzdjvpem</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210118034305/https://arxiv.org/pdf/2101.04927v1.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/28/87/28875ae28aedea9046c785c06ec8e2df3de3f4b3.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2101.04927v1" 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>

Image-to-Image Translation of Synthetic Samples for Rare Classes [article]

Edoardo Lanzini, Sara Beery
<span title="2021-06-23">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We use low-level feature alignment between source and target domains to make synthetic data for a rare species generated using a graphics engine more "realistic".  ...  One potential approach to increase the training data for these rare classes is to augment the limited real data with synthetic samples.  ...  Acknowledgements We would like to thank the USGS and NPS for providing data and Microsoft AI for Earth for providing compute resources.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2106.12212v1">arXiv:2106.12212v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/logankwygvcurkfh6eeybyatoq">fatcat:logankwygvcurkfh6eeybyatoq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210625130842/https://arxiv.org/pdf/2106.12212v1.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/f9/50/f95021b65dc2ff799a97f0ab0b46523a0956f490.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2106.12212v1" 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>

Generating All the Roads to Rome: Road Layout Randomization for Improved Road Marking Segmentation [article]

Tom Bruls, Horia Porav, Lars Kunze, Paul Newman
<span title="2019-07-10">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We demonstrate that training on these synthetic pairs improves mIoU of the segmentation of rare road marking classes during real-world deployment in complex urban environments by more than 12 percentage  ...  points, while performance for other classes is retained.  ...  We demonstrate quantitatively that training on these synthetic labels improves mIoU of the segmentation of rare road marking classes, for which it is expensive to attain sufficient real-world examples,  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1907.04569v1">arXiv:1907.04569v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/76nwbyj5fvbp5l3djnzlxgyd6m">fatcat:76nwbyj5fvbp5l3djnzlxgyd6m</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191222110928/https://arxiv.org/pdf/1907.04569v1.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/8a/d7/8ad77262c7ae5159264803eb71121c50d313550f.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1907.04569v1" 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>

Domain Adaptation for Rare Classes Augmented with Synthetic Samples [article]

Tuhin Das, Robert-Jan Bruintjes, Attila Lengyel, Jan van Gemert, Sara Beery
<span title="2021-10-23">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
While domain adaptation is generally applied on completely synthetic source domains and real target domains, we explore how domain adaptation can be applied when only a single rare class is augmented with  ...  To alleviate lower classification performance on rare classes in imbalanced datasets, a possible solution is to augment the underrepresented classes with synthetic samples.  ...  Acknowledgements Computational resources were provided by Microsoft AI for Earth.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2110.12216v1">arXiv:2110.12216v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/yob7ancqq5dk3av3oqkxuob4qi">fatcat:yob7ancqq5dk3av3oqkxuob4qi</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20211029065730/https://arxiv.org/pdf/2110.12216v1.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/c8/cb/c8cb1490c9d36286d27fb489355a81c9e470669e.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2110.12216v1" 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>

Integrating Expert Knowledge with Domain Adaptation for Unsupervised Fault Diagnosis [article]

Qin Wang, Cees Taal, Olga Fink
<span title="2021-07-05">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Data-driven fault diagnosis methods often require abundant labeled examples for each fault type.  ...  Experimental results demonstrate that the generated faults are effective for encoding fault type information and the domain adaptation is robust against the different levels of class imbalance between  ...  For example, for image generation tasks, Conditional Generative Adversarial Networks [45] concatenate the class vector with the feature vector to generate images conditioned on the single classes.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2107.01849v1">arXiv:2107.01849v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/hrz4pn73sfbvzpcy4ih5xr6n5m">fatcat:hrz4pn73sfbvzpcy4ih5xr6n5m</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210709084244/https://arxiv.org/pdf/2107.01849v1.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/f6/85/f685ea0810dd561e6bcb1dc814cbcd03df7fa28d.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2107.01849v1" 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>

Synthetic triphones from trajectory-based feature distributions

Jaco Badenhorst, Marelie H. Davel
<span title="">2015</span> <i title="IEEE"> 2015 Pattern Recognition Association of South Africa and Robotics and Mechatronics International Conference (PRASA-RobMech) </i> &nbsp;
We experiment with a new method to create synthetic models of rare and unseen triphones in order to supplement limited automatic speech recognition (ASR) training data.  ...  A trajectory model is used to characterise seen transitions at the spectral level, and these models are then used to create features for unseen or rare triphones.  ...  The results in Table III prove Table II for the number of triphones in each class.) Again, adding 10 examples does not provide the best improvement (79.31%).  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/robomech.2015.7359509">doi:10.1109/robomech.2015.7359509</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/tdebbjlwsfh5fjppj3ih4powim">fatcat:tdebbjlwsfh5fjppj3ih4powim</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20170921234815/https://researchspace.csir.co.za/dspace/bitstream/handle/10204/8737/Badenhorst_2015.pdf?sequence=1" 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/47/82/4782b12800b912bca4bd8efa76f5988712c05269.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/robomech.2015.7359509"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> ieee.com </button> </a>

The Impact of Local Data Characteristics on Learning from Imbalanced Data [chapter]

Jerzy Stefanowski
<span title="">2014</span> <i title="Springer International Publishing"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/2w3awgokqne6te4nvlofavy5a4" style="color: black;">Lecture Notes in Computer Science</a> </i> &nbsp;
     ) ( i i i new p n p x SMOTE limitations  "Blind" over-generalization in the directions of majority neighbors class  Number of synthetic examples -o - a global parameter  ...  select o of these neighbours (o -the amount of over-sampling desired)  Generate a synthetic example along the line between p and randomly selected example n  Can work with mixed attributes  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1007/978-3-319-08729-0_1">doi:10.1007/978-3-319-08729-0_1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/mxmijwylrrccpbkuy77jqrz34u">fatcat:mxmijwylrrccpbkuy77jqrz34u</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20170829073348/https://vid.ctb.upm.es/jrs2014/downloads/jstefanowski.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/b6/29/b62937818c22206231acd7557f2f10dc5e05c6a4.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1007/978-3-319-08729-0_1"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> springer.com </button> </a>

Multiple Classifier Prediction Improvements against Imbalanced Datasets through Added Synthetic Examples [chapter]

Herna L. Viktor, Hongyu Guo
<span title="">2004</span> <i title="Springer Berlin Heidelberg"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/2w3awgokqne6te4nvlofavy5a4" style="color: black;">Lecture Notes in Computer Science</a> </i> &nbsp;
In this way, the ensemble is able to focus not only on hard examples, but also on rare examples.  ...  In particular, the use of boosting which focuses on hard to learn examples, have application for difficult to learn problems.  ...  Firstly, we separately identify hard examples from and generate synthetic examples for different classes.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1007/978-3-540-27868-9_107">doi:10.1007/978-3-540-27868-9_107</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/mmqdytvir5eafk2yylyosuz5v4">fatcat:mmqdytvir5eafk2yylyosuz5v4</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20181030054526/https://link.springer.com/content/pdf/10.1007%2F978-3-540-27868-9_107.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/a0/f6/a0f624193f844449747043b071455c1396358a74.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1007/978-3-540-27868-9_107"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> springer.com </button> </a>

An Iterated Greedy Algorithm for Improving the Generation of Synthetic Patterns in Imbalanced Learning [chapter]

Francisco Javier Maestre-García, Carlos García-Martínez, María Pérez-Ortiz, Pedro Antonio Gutiérrez
<span title="">2017</span> <i title="Springer International Publishing"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/2w3awgokqne6te4nvlofavy5a4" style="color: black;">Lecture Notes in Computer Science</a> </i> &nbsp;
One of the most successful approaches for alleviating this problem is the generation of synthetic minority samples by convex combination of available ones.  ...  Within this framework, adaptive synthetic (ADASYN) sampling is a relatively new method which imposes weights on minority examples according to their learning complexity, in such a way that difficult examples  ...  boundary; cluster-oversampling [18] , which considers the so-called 'rare' regions, which are resampled individually; or safe-level SMOTE [2] and LN-SMOTE [24] , which generate new synthetic examples  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1007/978-3-319-59147-6_44">doi:10.1007/978-3-319-59147-6_44</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/emzs7dyfurd5fmzqdfx6hmopw4">fatcat:emzs7dyfurd5fmzqdfx6hmopw4</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20190428075425/http://discovery.ucl.ac.uk/10069046/1/Perez%20Ortiz_An_iterated_greedy_algorithm_for_improving_the_generation_of_synthetic_patterns_in_imbalanced_learning.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/00/12/0012231df42d4e09faa53b9e5e4b50116d728978.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1007/978-3-319-59147-6_44"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> springer.com </button> </a>

Exploring Rare Pose in Human Pose Estimation

Jihye Hwang, John Yang, Nojun Kwak
<span title="">2020</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;
FIGURE 2 : 2 An overall illustration of the synthetic rare pose data generation process. FIGURE 3 : 3 Examples of synthetic MPII rare pose data generated by the method shown inFig 2.  ...  Examples of synthetically generated pose data are shown in To generate more realistic synthesized samples, we have pre-trained a generator that translates styles from synthetic to real.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/access.2020.3033531">doi:10.1109/access.2020.3033531</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/oyhlzpqnyvdetnoa6gb233oyyy">fatcat:oyhlzpqnyvdetnoa6gb233oyyy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20201029132219/https://ieeexplore.ieee.org/ielx7/6287639/6514899/09239282.pdf?tp=&amp;arnumber=9239282&amp;isnumber=6514899&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/2c/f7/2cf7e8d20a063350c2b099cafa97252eef9142c3.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/access.2020.3033531"> <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>

DermGAN: Synthetic Generation of Clinical Skin Images with Pathology [article]

Amirata Ghorbani, Vivek Natarajan, David Coz, Yuan Liu
<span title="2019-11-20">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Finally, when using the synthetic images as a data augmentation technique for training a skin condition classifier, we observe that the model performs comparably to the baseline model overall while improving  ...  on rare but malignant conditions.  ...  Thanks also go to Yun Liu, Greg Corrado, and Erica Brand for their feedback on the project.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1911.08716v1">arXiv:1911.08716v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/ku5mgnk3vfgyzlg763o5np2m7a">fatcat:ku5mgnk3vfgyzlg763o5np2m7a</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200902053811/https://arxiv.org/pdf/1911.08716v1.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/34/a3/34a322efc82da0561ae8bbb14f958641bc7d9197.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1911.08716v1" 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>

Applying both positive and negative selection to supervised learning for anomaly detection

Xiaoshu Hang, Honghua Dai
<span title="">2005</span> <i title="ACM Press"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/fdhfwmjdwjbvxo6zc7cdt5hi7q" style="color: black;">Proceedings of the 2005 conference on Genetic and evolutionary computation - GECCO &#39;05</a> </i> &nbsp;
Two algorithms about synthetic generation of the anomaly class are proposed.  ...  It first learns the patterns of the normal class via co-evolutionary genetic algorithm, which is inspired from the positive selection, and then generates synthetic samples of the anomaly class, which is  ...  The advantage of the immunology-inspired synthetic generation of anomalous samples is that it is suitable for data sets with or without examples of the anomaly class.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1145/1068009.1068064">doi:10.1145/1068009.1068064</a> <a target="_blank" rel="external noopener" href="https://dblp.org/rec/conf/gecco/HangD05.html">dblp:conf/gecco/HangD05</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/i5duwiwsifgeraglrfqpd3kjki">fatcat:i5duwiwsifgeraglrfqpd3kjki</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20170705171423/http://www.cs.bham.ac.uk/~wbl/biblio/gecco2005/docs/p345.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/73/dc/73dcd1aac1d573ad8e805ffa8eebd3f1ba95b0ae.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1145/1068009.1068064"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> acm.org </button> </a>

SMOTEBoost: Improving Prediction of the Minority Class in Boosting [chapter]

Nitesh V. Chawla, Aleksandar Lazarevic, Lawrence O. Hall, Kevin W. Bowyer
<span title="">2003</span> <i title="Springer Berlin Heidelberg"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/2w3awgokqne6te4nvlofavy5a4" style="color: black;">Lecture Notes in Computer Science</a> </i> &nbsp;
Unlike standard boosting where all misclassified examples are given equal weights, SMOTEBoost creates synthetic examples from the rare or minority class, thus indirectly changing the updating weights and  ...  SMOTE (Synthetic Minority Over-sampling TEchnique) is specifically designed for learning from imbalanced data sets.  ...  We also thank Philip Kegelmeyer for his helpful feedback. We would also like to thank anonymous reviewers for their useful comments on the paper.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1007/978-3-540-39804-2_12">doi:10.1007/978-3-540-39804-2_12</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/aejdrb7s7fcwpiwnowz6aczr74">fatcat:aejdrb7s7fcwpiwnowz6aczr74</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20170829153335/https://www3.nd.edu/~dial/publications/chawla2003smoteboost.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/99/2f/992fae89bbff1079503b6d0e4e1e6f2ef6d24ced.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1007/978-3-540-39804-2_12"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> springer.com </button> </a>
&laquo; Previous Showing results 1 &mdash; 15 out of 104,450 results