Filters








631,911 Hits in 3.9 sec

Three scenarios for continual learning [article]

Gido M. van de Ven, Andreas S. Tolias
<span title="2019-04-15">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Using the split and permuted MNIST task protocols, for each scenario we carry out an extensive comparison of recently proposed continual learning methods.  ...  To enable more structured comparisons, we describe three continual learning scenarios based on whether at test time task identity is provided and--in case it is not--whether it must be inferred.  ...  Acknowledgments We thank Mengye Ren, Zhe Li and anonymous reviewers for comments on various parts of this work.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1904.07734v1">arXiv:1904.07734v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/r7nvzhzt2nd27bucsm7uopwwga">fatcat:r7nvzhzt2nd27bucsm7uopwwga</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200825132124/https://arxiv.org/pdf/1904.07734v1.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/c5/d0/c5d0c9ba5c783540cdfa3969b7fd2c853c9d2dc8.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1904.07734v1" 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>

CLRS: Continual Learning Benchmark for Remote Sensing Image Scene Classification

Haifeng Li, Hao Jiang, Xin Gu, Jian Peng, Wenbo Li, Liang Hong, Chao Tao
<span title="2020-02-24">2020</span> <i title="MDPI AG"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/taedaf6aozg7vitz5dpgkojane" style="color: black;">Sensors</a> </i> &nbsp;
Finally, several mainstream continual learning methods are tested and analyzed under three continual learning scenarios, and the results can be used as a baseline for future work.  ...  First, this study gives the criteria for training batches that have been partitioned into three continual learning scenarios, and proposes a large-scale remote sensing image scene classification database  ...  (4) We tested and analyzed several mainstream continual learning methods for three continual scenarios, and the results can be used as a baseline for future work.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.3390/s20041226">doi:10.3390/s20041226</a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pubmed/32102294">pmid:32102294</a> <a target="_blank" rel="external noopener" href="https://pubmed.ncbi.nlm.nih.gov/PMC7070946/">pmcid:PMC7070946</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/ysbpcwaxjbhtxcwd7m3lrlrie4">fatcat:ysbpcwaxjbhtxcwd7m3lrlrie4</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200227232529/https://res.mdpi.com/d_attachment/sensors/sensors-20-01226/article_deploy/sensors-20-01226.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/f9/17/f917e2e5cded45e3d886d618265ac693791c22e0.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.3390/s20041226"> <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> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7070946" title="pubmed link"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> pubmed.gov </button> </a>

Adaptive Explainable Continual Learning Framework for Regression Problems with Focus on Power Forecasts [article]

Yujiang He
<span title="2022-02-07">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In this article, two continual learning scenarios will be proposed to describe the potential challenges in this context.  ...  Besides, based on our previous work regarding the CLeaR framework, which is short for continual learning for regression tasks, the work will be further developed to enable models to extend themselves and  ...  Bernhard Sick and supported within the Digital-Twin-Solar (03EI6024E) project, funded by BMWi: Deutsches Bundesministerium für Wirtschaft und Energie/German Federal Ministry for Economic Affairs and Energy  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2108.10781v2">arXiv:2108.10781v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/zfylub527rdkvp43iifowwcoca">fatcat:zfylub527rdkvp43iifowwcoca</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220211001625/https://arxiv.org/pdf/2108.10781v2.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/72/da7285a34897cf088e63cc25d8f34308fb5104b9.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2108.10781v2" 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>

Re-evaluating Continual Learning Scenarios: A Categorization and Case for Strong Baselines [article]

Yen-Chang Hsu, Yen-Cheng Liu, Anita Ramasamy, Zsolt Kira
<span title="2019-01-23">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We conclude with several suggestions for creating harder evaluation scenarios and future research directions. The code is available at https://github.com/GT-RIPL/Continual-Learning-Benchmark  ...  Continual learning has received a great deal of attention recently with several approaches being proposed.  ...  Acknowledgments This research is supported by DARPA's Lifelong Learning Machines (L2M) program, under Cooperative Agreement HR0011-18-2-001.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1810.12488v4">arXiv:1810.12488v4</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/gyel42dzhfah5acsmmkohprale">fatcat:gyel42dzhfah5acsmmkohprale</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200829101935/https://arxiv.org/pdf/1810.12488v4.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/9f/bf/9fbf092b5109ecc24f4d246e6ced8301b644ef91.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1810.12488v4" 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>

Are patient-centered care values as reflected in teaching scenarios really being taught when implemented by teaching faculty? A discourse analysis on an Indonesian medical school's curriculum

Mora Claramita, Adi H Sutomo, Mark A Graber, Albert JJ Scherpbier
<span title="">2011</span> <i title="Springer Nature"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/lkbejn6a6nbshebsjawf2rf254" style="color: black;">Asia Pacific Family Medicine</a> </i> &nbsp;
Methods: Discourse analysis was done by 3 general practitioners of scenarios and learning objectives of an Indonesian undergraduate medical curriculum.  ...  Most sentences in the scenarios were more relevant to patient-centered care while most sentences in the learning objectives were more inclined towards disease-perspectives.  ...  this work, doctor Manik Kharismayekti for her commitment towards primary care practice and the NPT Project UGM-UM for a great and continuous support.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1186/1447-056x-10-4">doi:10.1186/1447-056x-10-4</a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pubmed/21513582">pmid:21513582</a> <a target="_blank" rel="external noopener" href="https://pubmed.ncbi.nlm.nih.gov/PMC3111348/">pmcid:PMC3111348</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/3acbsjpwd5aqpoaalfeqibxsfi">fatcat:3acbsjpwd5aqpoaalfeqibxsfi</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20140519172938/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3111348/pdf/1447-056X-10-4.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/f9/81/f98172a2c4ea6d913f3f535797aaedb9f9d7d4b3.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1186/1447-056x-10-4"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="unlock alternate icon" style="background-color: #fb971f;"></i> springer.com </button> </a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3111348" title="pubmed link"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> pubmed.gov </button> </a>

Synchronizing Asynchronous Collaborative Learners [chapter]

Johan Lundin
<span title="">2003</span> <i title="Springer Netherlands"> Communities and Technologies </i> &nbsp;
The three methods were implemented and evaluated in an organization using a Net-scenario, the Net-scenario as a system and a methodology based on role-playing to initiate collaborative learning.  ...  To address this problem three methods for synchronization (synchronization points) are suggested: locked scenes, written instruction and collaborative production.  ...  Acknowledgments This project was made possible through the financial support from SITI (The Swedish Research Institute for Information Technology), the collaboration with Farshad Taghizadeh and ADBkontoret  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1007/978-94-017-0115-0_22">doi:10.1007/978-94-017-0115-0_22</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/u5hgohcet5d5lklc67ozzf67mq">fatcat:u5hgohcet5d5lklc67ozzf67mq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20070710143743/http://www.iisi.de:80/fileadmin/IISI/upload/C_T/2003/Lundin.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/3b/25/3b25914803deedd223755dce0d218d4b67270714.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1007/978-94-017-0115-0_22"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> springer.com </button> </a>

Ex-Model: Continual Learning from a Stream of Trained Models [article]

Antonio Carta, Andrea Cossu, Vincenzo Lomonaco, Davide Bacciu
<span title="2021-12-13">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We further contribute with three ex-model continual learning algorithms and an empirical setting comprising three datasets (MNIST, CIFAR-10 and CORe50), and eight scenarios, where the proposed algorithms  ...  Being able to learn, adapt, and generalize continually in an efficient, effective, and scalable way is fundamental for a sustainable development of Artificial Intelligent systems.  ...  Stream accuracy computed on the test set for CORe50 continual learning scenarios.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2112.06511v1">arXiv:2112.06511v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/cp7fc45ehzbxfa7risbu56a2bi">fatcat:cp7fc45ehzbxfa7risbu56a2bi</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20211215003713/https://arxiv.org/pdf/2112.06511v1.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/e8/c7/e8c7d901cf569d6e224591494e184b20827350db.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2112.06511v1" 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>

Continuum: Simple Management of Complex Continual Learning Scenarios [article]

Arthur Douillard, Timothée Lesort
<span title="2021-02-11">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
One implementation difficulty in these field is to create data loaders that simulate non-iid scenarios. Indeed, data loaders are a key component for continual algorithms.  ...  Continual learning is a machine learning sub-field specialized in settings with non-iid data. Hence, the training data distribution is not static and drifts through time.  ...  Organization To create continual learning scenarios, Continuum decomposes the data management into three levels of data 1 Code available at https : / / github . com / Continvvm / continuum.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2102.06253v1">arXiv:2102.06253v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/v337xxdjkrhmlowepbisrdfvjq">fatcat:v337xxdjkrhmlowepbisrdfvjq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210216090102/https://arxiv.org/pdf/2102.06253v1.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/b9/41/b94122b1f72fab95e325c4f18e7efa37781e8c17.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2102.06253v1" 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>

Regularization Shortcomings for Continual Learning [article]

Timothée Lesort, Andrei Stoian, David Filliat
<span title="2021-04-04">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Moreover, we show that it can have some important consequences on continual multi-tasks reinforcement learning or in pre-trained models used for continual learning.  ...  : the class-incremental scenario.  ...  Thanks also to Vyshakh Palli Thazha for proofreading this article.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1912.03049v4">arXiv:1912.03049v4</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/iurk4btdi5d5xfgzrx5xi6pvuq">fatcat:iurk4btdi5d5xfgzrx5xi6pvuq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210407001550/https://arxiv.org/pdf/1912.03049v4.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/85/68/856810aa07610306fe2da1f44b5efa77dfbfd4dc.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1912.03049v4" 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 lessons from sunk costs

Brian H. Bornstein, Gretchen B. Chapman
<span title="">1995</span> <i title="American Psychological Association (APA)"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/gyez7a2i2rbcxlwgkbxo7ztji4" style="color: black;">Journal of experimental psychology. Applied</a> </i> &nbsp;
cello Study participants rated the quality of several arguments for continuing an original plan in sunk cost situations in order to (a) avoid wasting resources, (b) learn to make better decisions, (c)  ...  An obvious prediction that is consistent with the sunk cost effect is that the greater the initial investment, the stronger the effect should be; for example, one should be more likely to continue taking  ...  for continuing the original plan: waste, learn-a-lesson, consistency, and punishment.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1037/1076-898x.1.4.251">doi:10.1037/1076-898x.1.4.251</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/4jqhj32gqjhabdgjlq2lm4wm4m">fatcat:4jqhj32gqjhabdgjlq2lm4wm4m</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191224110007/http://digitalcommons.unl.edu/cgi/viewcontent.cgi?article=1293&amp;context=psychfacpub" 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/5c/dc/5cdc7b48e2d8cad3590a24f498435a0fefa6261f.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1037/1076-898x.1.4.251"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> Publisher / doi.org </button> </a>

Technical Report for ICCV 2021 Challenge SSLAD-Track3B: Transformers Are Better Continual Learners [article]

Duo Li, Guimei Cao, Yunlu Xu, Zhanzhan Cheng, Yi Niu
<span title="2022-01-13">2022</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In the SSLAD-Track 3B challenge on continual learning, we propose the method of COntinual Learning with Transformer (COLT).  ...  In this report, we first introduce the overall framework of continual learning for object detection. Then, we analyse the key elements' effect on withstanding catastrophic forgetting in our solution.  ...  For continual learning on the SSLAD dataset, we train the model on 8 GPUs, one sample for each GPU.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2201.04924v1">arXiv:2201.04924v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/lalu5sa5szgdrewy2ebyjkmagq">fatcat:lalu5sa5szgdrewy2ebyjkmagq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220120132650/https://arxiv.org/pdf/2201.04924v1.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/d4/7e/d47e71698b5472bc0f17a8e304317c16f1acd078.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2201.04924v1" 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>

What is Wrong with Continual Learning in Medical Image Segmentation? [article]

Camila Gonzalez, Georgios Sakas, Anirban Mukhopadhyay
<span title="2020-10-21">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Many techniques exist to learn continuously for classification tasks, and several have been adapted to semantic segmentation.  ...  We show that the benchmark outperforms two popular continual learning methods for the task of T2-weighted MR prostate segmentation.  ...  Three continual learning scenarios, exemplified for the case of prostate segmentation.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2010.11008v1">arXiv:2010.11008v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/ddypzmpmzjarpoqn5u6bk3zpcu">fatcat:ddypzmpmzjarpoqn5u6bk3zpcu</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20201025170941/https://arxiv.org/pdf/2010.11008v1.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/08/f6/08f6be35d7d4212352c60bda38f90201d978f391.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2010.11008v1" 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>

Batch-level Experience Replay with Review for Continual Learning [article]

Zheda Mai, Hyunwoo Kim, Jihwan Jeong, Scott Sanner
<span title="2020-07-11">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
different continual learning scenarios.  ...  The CVPR 2020 CLVision Continual Learning for Computer Vision challenge is dedicated to evaluating and advancing the current state-of-the-art continual learning methods using the CORe50 dataset with three  ...  Conclusion In this paper, we described our approach for the CVPR 2020 CLVision Continual Learning for Computer Vision challenge.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2007.05683v1">arXiv:2007.05683v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/2llu6ex7dnadvfkci3nk7ehwti">fatcat:2llu6ex7dnadvfkci3nk7ehwti</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200908100941/https://arxiv.org/pdf/2007.05683v1.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/58/39/5839f897278478c8b6aa268e9281078c393f0236.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2007.05683v1" 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>

New Metrics and Experimental Paradigms for Continual Learning

Tyler L. Hayes, Ronald Kemker, Nathan D. Cahill, Christopher Kanan
<span title="">2018</span> <i title="IEEE"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/ilwxppn4d5hizekyd3ndvy2mii" style="color: black;">2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)</a> </i> &nbsp;
In this abstract, we formalize the continual learning paradigm and propose new benchmarks for assessing continual learning agents.  ...  In order for a robotic agent to learn successfully in an uncontrolled environment, it must be able to immediately alter its behavior.  ...  While this is not a realistic scenario for a robot, this scenario should be the easiest for a continual learning model to rival an offline learner.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/cvprw.2018.00273">doi:10.1109/cvprw.2018.00273</a> <a target="_blank" rel="external noopener" href="https://dblp.org/rec/conf/cvpr/HayesKCK18.html">dblp:conf/cvpr/HayesKCK18</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/25j6v7hmzfcalie6vov5dlukvq">fatcat:25j6v7hmzfcalie6vov5dlukvq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200318215255/http://openaccess.thecvf.com/content_cvpr_2018_workshops/papers/w40/Hayes_New_Metrics_and_CVPR_2018_paper.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/55/99/559931c910cba96690ce71026c51969c997081e5.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/cvprw.2018.00273"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> ieee.com </button> </a>

Life-Long Multi-Task Learning of Adaptive Path Tracking Policy for Autonomous Vehicle [article]

Cheng Gong and Jianwei Gong and Chao Lu and Zhe Liu and Zirui Li
<span title="2021-09-15">2021</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Thirdly, a memory evaluation and updating method is applied to optimize memory structure for life-long learning which enables the policy to learn toward selected directions.  ...  Thus, with continual multi-task knowledge learned, the policy can iteratively adapt to new tasks and improve its performance with knowledge from new tasks.  ...  Figure 4 . 4 Max lateral deviation of executing path tracking with three learned policies in test scenario S1.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2109.07210v1">arXiv:2109.07210v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/qf33vl6lczehnimewc2ygh53x4">fatcat:qf33vl6lczehnimewc2ygh53x4</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210921130024/https://arxiv.org/ftp/arxiv/papers/2109/2109.07210.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/0d/fb/0dfbf7e324c6f690322e4e16ded31f53602b1e5c.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2109.07210v1" 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 631,911 results