Filters








5,544 Hits in 7.5 sec

Active Learning With Sampling by Uncertainty and Density for Data Annotations

J Zhu, H Wang, B K Tsou, M Ma
2010 IEEE Transactions on Audio, Speech, and Language Processing  
Index Terms-Active learning, density-based re-ranking, sampling by uncertainty and density, text classification, uncertainty sampling, word sense disambiguation (WSD).  ...  Experimental results of active learning for word sense disambiguation and text classification tasks using six real-world evaluation data sets demonstrate the effectiveness of the proposed methods.  ...  Fig. 6 . 6 Results of margin-based uncertainty sampling and density-based re-ranking methods for active learning with SVMs on the Interest data set.  ... 
doi:10.1109/tasl.2009.2033421 fatcat:i5l6yq7i5rbg7o3itozm2e7sua

Active Learning from Positive and Unlabeled Data

Alireza Ghasemi, Hamid R. Rabiee, Mohsen Fadaee, Mohammad T. Manzuri, Mohammad H. Rohban
2011 2011 IEEE 11th International Conference on Data Mining Workshops  
In this paper we propose an active learning algorithm that can work when only samples of one class as well as a set of unlabelled data are available.  ...  Such problems arise in many real-world situations and are known as the problem of learning from positive and unlabeled data.  ...  ACKNOWLEDGEMENT The authors would like to thank the AICTC Research Center for supporting this work.  ... 
doi:10.1109/icdmw.2011.20 dblp:conf/icdm/GhasemiRFMR11 fatcat:ku5zxgnlm5ccbkbyle5fpxubpq

Uncertainty-based active learning with instability estimation for text classification

Jingbo Zhu, Matthew Ma
2012 ACM Transactions on Speech and Language Processing  
ACM Reference Format: Zhu, J. and Ma, M. 2012. Uncertainty-Based active learning with instability estimation for text classification.  ...  and density methods achieve better effectiveness in annotation cost reduction than random sampling and traditional entropy-based uncertainty sampling.  ...  ACKNOWLEDGMENTS Thanks to Muhua Zhu for implementing some comparison experiments of active learning with SVMs.  ... 
doi:10.1145/2093153.2093154 dblp:journals/tslp/ZhuM12 fatcat:lfimankrmjdbzlhx3cis5aytoe

Adaptive Active Learning for Image Classification

Xin Li, Yuhong Guo
2013 2013 IEEE Conference on Computer Vision and Pattern Recognition  
In this paper, we present a novel adaptive active learning approach that combines an information density measure and a most uncertainty measure together to select critical instances to label for image  ...  Recently active learning has attracted a lot of attention in computer vision field, as it is time and cost consuming to prepare a good set of labeled images for vision data analysis.  ...  The methods in [19, 32] employ the unlabeled data by using the prior density p(x) as weights for uncertainty measures.  ... 
doi:10.1109/cvpr.2013.116 dblp:conf/cvpr/LiG13 fatcat:peulo4vwi5afnnvjytwrbf7vfe

CRF-based active learning for Chinese named entity recognition

Lin Yao, Chengjie Sun, Shaofeng Li, Xiaolong Wang, Xuan Wang
2009 2009 IEEE International Conference on Systems, Man and Cybernetics  
This paper proposes a new active learning strategy based on Information Density (ID) integrated with CRFs for Chinese Named Entity Recognition (NER).  ...  Based on different query strategies, active learning can combine with other machine learning methods to reduce the annotation cost while maintaining the accuracy.  ...  Active Learning Query Strategies Active Learning methods rely on different strategies for sampling unlabeled instances.  ... 
doi:10.1109/icsmc.2009.5346315 dblp:conf/smc/YaoSLWW09 fatcat:vrsj4bvn7bdkdblsuk5z4ujkma

Exploring Representativeness and Informativeness for Active Learning

Bo Du, Zengmao Wang, Lefei Zhang, Liangpei Zhang, Wei Liu, Jialie Shen, Dacheng Tao
2017 IEEE Transactions on Cybernetics  
Although combining representativeness and informativeness of samples has been proven promising for active sampling, state-of-the-art methods perform well under certain data structures.  ...  Then can we find a way to fuse the two active sampling criteria without any assumption on data? This paper proposes a general active learning framework that effectively fuses the two criteria.  ...  Proposed Active Learning Method Based on the proposed optimal framework, we propose an active learning method relying on the probability estimates of class membership for all the samples. 1) Computing  ... 
doi:10.1109/tcyb.2015.2496974 pmid:26595936 fatcat:fti4vegaavfw7o4wqe6lkatywm

Exploring Representativeness and Informativeness for Active Learning [article]

Bo Du, Zengmao Wang, Lefei Zhang, Liangpei Zhang, Wei Liu, Jialie Shen, Dacheng Tao
2019 arXiv   pre-print
Although combining representativeness and informativeness of samples has been proven promising for active sampling, state-of-the-art methods perform well under certain data structures.  ...  Then can we find a way to fuse the two active sampling criteria without any assumption on data? This paper proposes a general active learning framework that effectively fuses the two criteria.  ...  ACKNOWLEDGMENT The authors would like to thank the handing editor and the anonymous reviewers for their careful reading and helpful remarks, which have contributed in improving the quality of this paper  ... 
arXiv:1904.06685v1 fatcat:r6rafqpcojdkreoicgqvmn4jse

Shift-Pessimistic Active Learning Using Robust Bias-Aware Prediction

Anqi Liu, Lev Reyzin, Brian Ziebart
2015 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
Existing approaches to active learning are generally optimistic about their certainty with respect to data shift between labeled and unlabeled data.  ...  We investigate the theoretical benefits of this approach and demonstrate its empirical advantages on probabilistic binary classification tasks.  ...  Acknowledgments This material is based upon work supported by the National Science Foundation under Grant No. #1227495.  ... 
doi:10.1609/aaai.v29i1.9609 fatcat:qksmjpqiqfdqdpbom7qcc2h63a

An Active Learning Method Based on Variational Autoencoder and DBSCAN Clustering

Fang Chen, Tao Zhang, Ruilin Liu, Ricardo Aler
2021 Computational Intelligence and Neuroscience  
Active learning is aimed to sample the most informative data from the unlabeled pool, and diverse clustering methods have been applied to it.  ...  In this paper, we propose a new active learning method combined with variational autoencoder (VAE) and density-based spatial clustering of applications with noise (DBSCAN).  ...  Later, Gissin and Shai proposed a discriminative active learning method [12] in which an uncertainty idea is also used to sample the unlabeled data with the top-K highest possibility when the batch size  ... 
doi:10.1155/2021/9952596 pmid:34381500 pmcid:PMC8352707 fatcat:y65lxa4u4bccni2v77uax5g2tm

From Active to Proactive Learning Methods [chapter]

Pinar Donmez, Jaime G. Carbonell
2010 Studies in Computational Intelligence  
We group active learning methods into three major categories based on their underlying sampling criteria; namely, 1) uncertainty-based sampling strategies, 2) density-based strategies and 3) ensemble methods  ...  The reason is that density based methods sample from maximal-density unlabeled regions, and thus help establish the initial classification boundary where it affects the most remaining unlabeled data.  ... 
doi:10.1007/978-3-642-05177-7_5 fatcat:memkyn4yxzdyrp3cc2ex727nzy

Applying active learning to assertion classification of concepts in clinical text

Yukun Chen, Subramani Mani, Hua Xu
2012 Journal of Biomedical Informatics  
The outcome is reported in the global ALC score, based on the Area under the average Learning Curve of the AUC (Area Under the Curve) score.  ...  We implemented several existing and newly developed active learning algorithms and assessed their uses.  ...  Acknowledgments This study was supported in part by NIH Grants NLM R01LM010681 and NCI R01CA141307. The datasets used were ob-  ... 
doi:10.1016/j.jbi.2011.11.003 pmid:22127105 pmcid:PMC3306548 fatcat:kyyvw5b5ufectmsrc4yisyd5e4

Asymmetric propagation based batch mode active learning for image retrieval

Biao Niu, Jian Cheng, Xiao Bai, Hanqing Lu
2013 Signal Processing  
and density into batch mode active learning in relevance feedback.  ...  In order to alleviate the burden of labeling, active learning method has been introduced to select the most informative samples for labeling.  ...  Acknowledgments This work was supported in part by the 973 Programm under Project 2010CB327905, by the National Natural Science Foundation of China under Grant nos. 61170127, 60975010, 60833006, and 61070104  ... 
doi:10.1016/j.sigpro.2012.07.018 fatcat:vdzrdjqsbjadth2igniaqqdxxi

Improving Relevance Feedback for Image Retrieval with Asymmetric Sampling

Biao Niu, Jian Cheng, Hanqing Lu
2012 2012 IEEE International Conference on Multimedia and Expo  
In this paper, we presents a novel batch mode active learning method for informative sample selection.  ...  Inspired by graph propagation, we consider the certainty of labels as asymmetric propagation information on graph, and formulate the correlation between labeled samples and unlabeled samples in an united  ...  In this paper, we present a novel batch mode active learning based on asymmetric propagation by modeling the criteria uncertainty, diversity, and density with the selection scheme.  ... 
doi:10.1109/icme.2012.127 dblp:conf/icmcs/NiuCL12 fatcat:l37w2hyqknbzzhskrziwhtn2ga

Selecting Influential Examples: Active Learning with Expected Model Output Changes [chapter]

Alexander Freytag, Erik Rodner, Joachim Denzler
2014 Lecture Notes in Computer Science  
This results in a score for each unlabeled example that can be used for active learning with a broad range of models and learning algorithms.  ...  The key idea of our approach is to measure the expected change of model outputs, a concept that generalizes previous methods based on expected model change and incorporates the underlying data distribution  ...  We show in our experiments that this modification improves the performance of previous active learning methods and therefore also offers a fairer comparison to our new active learning methods based on  ... 
doi:10.1007/978-3-319-10593-2_37 fatcat:6gtxpf4hsfc75a6lammpkxjlee

Ask me better questions

Parisa Rashidi, Diane J. Cook
2011 Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining - KDD '11  
Most traditional active learning methods pose a very specific query to the oracle, i.e. they ask for the label of an unlabeled example.  ...  It can construct generic active learning queries based on rule induction from multiple unlabeled instances.  ...  Active learning methods usually select an informative unlabeled instance and ask the oracle for the label of the instance.  ... 
doi:10.1145/2020408.2020559 dblp:conf/kdd/RashidiC11 fatcat:allcktjz4vcxhiyexlgjepybla
« Previous Showing results 1 — 15 out of 5,544 results