Filters








1,854 Hits in 4.5 sec

Convex Formulation of Multiple Instance Learning from Positive and Unlabeled Bags [article]

Han Bao, Tomoya Sakai, Issei Sato, Masashi Sugiyama
2018 arXiv   pre-print
A learning framework called PU learning (positive and unlabeled learning) can address this problem. In this paper, we propose a convex PU learning method to solve an MIL problem.  ...  Multiple instance learning (MIL) is a variation of traditional supervised learning problems where data (referred to as bags) are composed of sub-elements (referred to as instances) and only bag labels  ...  Multiple Instance Learning from Positive and Unlabeled Bags We formulate the problem of multiple instance learning from positive and unlabeled bags (PU-MIL).  ... 
arXiv:1704.06767v3 fatcat:q6lmjvhuj5e35dio64ivpiqtli

What's it going to cost you?: Predicting effort vs. informativeness for multi-label image annotations

Sudheendra Vijayanarasimhan, Kristen Grauman
2009 2009 IEEE Conference on Computer Vision and Pattern Recognition  
However, for visual category learning, the active selection problem is particularly complex: a single image will typically contain multiple object labels, and an annotator could provide multiple types  ...  We present an active learning framework that predicts the tradeoff between the effort and information gain associated with a candidate image annotation, thereby ranking unlabeled and partially labeled  ...  In MIL [4] , the learner is given sets (bags) of instances and told that at least one example from a positive bag is positive, while none of the members in a negative bag is positive.  ... 
doi:10.1109/cvpr.2009.5206705 dblp:conf/cvpr/VijayanarasimhanG09 fatcat:bnhvhsunifd67hdpg53qrwbks4

What's it going to cost you?: Predicting effort vs. informativeness for multi-label image annotations

S. Vijayanarasimhan, K. Grauman
2009 2009 IEEE Conference on Computer Vision and Pattern Recognition  
However, for visual category learning, the active selection problem is particularly complex: a single image will typically contain multiple object labels, and an annotator could provide multiple types  ...  We present an active learning framework that predicts the tradeoff between the effort and information gain associated with a candidate image annotation, thereby ranking unlabeled and partially labeled  ...  In MIL [4] , the learner is given sets (bags) of instances and told that at least one example from a positive bag is positive, while none of the members in a negative bag is positive.  ... 
doi:10.1109/cvprw.2009.5206705 fatcat:nlgfala2prce7befigozpvtjku

Learning Graph Neural Networks with Positive and Unlabeled Nodes [article]

Man Wu, Shirui Pan, Lan Du, Xingquan Zhu
2021 arXiv   pre-print
Two novel risk estimators are further employed to aggregate long-short-distance networks, for PU learning and the loss is back-propagated for model learning.  ...  To enable graph neural network learning, existing works typically assume that labeled nodes, from two or multiple classes, are provided, so that a discriminative classifier can be learned from the labeled  ...  IIS-1763452, CNS-1828181, and IIS-2027339.  ... 
arXiv:2103.04683v1 fatcat:7fofa7qkd5ggfdusspij5axwvu

Batch mode Adaptive Multiple Instance Learning for computer vision tasks

Wen Li, Lixin Duan, I. W. Tsang, Dong Xu
2012 2012 IEEE Conference on Computer Vision and Pattern Recognition  
Multiple Instance Learning (MIL) has been widely exploited in many computer vision tasks, such as image retrieval, object tracking and so on.  ...  In this paper, we propose a novel batch mode framework, namely Batch mode Adaptive Multiple Instance Learning (BAMIL), to accelerate the instance-level MIL methods.  ...  Acknowledgement This research is supported by the Singapore National Research Foundation under its Interactive & Digital Media (IDM) Public Sector R&D Funding Initiative and administered by the IDM Programme  ... 
doi:10.1109/cvpr.2012.6247949 dblp:conf/cvpr/LiDTX12 fatcat:5lfkmrwcxvee5pyylpofxi36be

Explainable Fraud Detection for Few Labeled Time Series Data

Zhiwen Xiao, Jianbin Jiao, Liguo Zhang
2021 Security and Communication Networks  
The main contribution of our work is to propose an explainable classification method in the framework of multiple instance learning (MIL), which incorporates the AP clustering method in the self-training  ...  fraud losses and management costs.  ...  Our main contribution is to propose an explainable classification method by improving the multiple instance learning (MIL) framework so as to realize fraud detection for time series data with few labeled  ... 
doi:10.1155/2021/9941464 fatcat:ilijubow2fgczjx7l7og6z65ma

A Novel Learning Formulation in a unified Min-Max Framework for Computer Aided Diagnosis

Sree Kanth
2013 IOSR Journal of Computer Engineering  
A bag is labeled positive if at least one instance in it is positive and negative if all the instances in it are negative. The www.iosrjournals.org  ...  To address all these problems, we propose a novel learning formulation to combine cascade classification and multiple instance learning (MIL) in a unified min-max framework, leading to a joint optimization  ...  , which is different from a standard MIL problem where instances in a positive bag can be unlabeled or even negative.  ... 
doi:10.9790/0661-1344452 fatcat:x3ufrrdjpff4pf65uh4vrvggsi

A high-performance semi-supervised learning method for text chunking

Rie Kubota Ando, Tong Zhang
2005 Proceedings of the 43rd Annual Meeting on Association for Computational Linguistics - ACL '05  
The idea is to find "what good classifiers are like" by learning from thousands of automatically generated auxiliary classification problems on unlabeled data.  ...  In machine learning, whether one can build a more accurate classifier by using unlabeled data (semi-supervised learning) is an important issue.  ...  Create training data ´ µ for each auxiliary problem from unlabeled data . 2. Compute ¢ from through SVD-ASO. 3. Minimize the empirical risk on the labeled data: (1) .  ... 
doi:10.3115/1219840.1219841 dblp:conf/acl/AndoZ05 fatcat:b222o7nponbn3dqifbdci5ng2i

Evaluating the Performance Estimators via Machine Learning Supervised Learning Algorithms for Dataset Threshold

Warda Imtiaz, Humaraia Abdul Ghafoor, Rabeea Sehar, Tahira Mahboob, Memoona Khanum
2015 International Journal of Computer Applications  
We focus on semisurprised framework which incorporates labeled and unlabeled data in the general-purpose learner.  ...  Represented theorems provide the theoretical base for algorithms. 2.2) Unsupervised and Supervised Machine Learning in User Modeling for Intelligent Learning Environments (S.Amershi and C.Conati) Framework  ...  In supervised learning risk minimization is presented as the suitable criteria to optimize. Basically supervised learning is used for implement risk minimization.  ... 
doi:10.5120/21132-4059 fatcat:aozpcysea5d2hoh5aazftp6xjq

Not-so-supervised: a survey of semi-supervised, multi-instance, and transfer learning in medical image analysis

Veronika Cheplygina, Marleen de Bruijne, Josien P.W. Pluim
2019 Medical Image Analysis  
We give an overview of semi-supervised, multiple instance, and transfer learning in medical imaging, both in diagnosis or segmentation tasks.  ...  We also discuss connections between these learning scenarios, and opportunities for future research.  ...  Ilse, Ragav Venkatesan and Wouter Kouw.  ... 
doi:10.1016/j.media.2019.03.009 pmid:30959445 fatcat:bbgz7v3ixvggnksuvxsxernobm

Learning From Positive and Unlabeled Data: A Survey [article]

Jessa Bekker, Jesse Davis
2018 arXiv   pre-print
Learning from positive and unlabeled data or PU learning is the setting where a learner only has access to positive examples and unlabeled data.  ...  The assumption is that the unlabeled data can contain both positive and negative examples.  ...  In an empirical risk minimization framework, this means finding the classifier g that minimizes the risk, given some loss function L R(g) = αE f+ L + (g(x)) + (1 − α)E f− L − (g(x)) , where L + (ŷ) and  ... 
arXiv:1811.04820v1 fatcat:2qi5g4xke5fljpw2iz7nhudc5a

Not-so-supervised: a survey of semi-supervised, multi-instance, and transfer learning in medical image analysis [article]

Veronika Cheplygina, Marleen de Bruijne, Josien P. W. Pluim
2018 arXiv   pre-print
We review semi-supervised, multiple instance, and transfer learning in medical imaging, both in diagnosis/detection or segmentation tasks.  ...  We also discuss connections between these learning scenarios, and opportunities for future research.  ...  The idea is to learn from only positive and unlabeled examples, which may happen when during annotation, the expert can miss other positives.  ... 
arXiv:1804.06353v2 fatcat:xke66fsrgjanjmopwti5tze4fm

Progressive Learning for Interactive Surveillance Scenes Retrieval

Jerome Meessen, Xavier Desurmont, Jean-Francois Delaigle, Christophe De Vleeschouwer, Benoit Macq
2007 2007 IEEE Conference on Computer Vision and Pattern Recognition  
The proposed method is based on very low-cost features extraction and integrates relevance feedback, multiple-instance SVM classification and active learning.  ...  Repeatable experiments on both simulated and real data demonstrate the efficiency of the approach and show how it allows reaching high retrieval performances.  ...  Inspired by [2] , we propose a method for guiding the crossvalidation procedure in our multiple-instance framework, derived from the 1-norm formulation.  ... 
doi:10.1109/cvpr.2007.383517 dblp:conf/cvpr/MeessenDDVM07 fatcat:ktdx32mucjdqpehz5wx2o7qzqu

A min-max framework of cascaded classifier with multiple instance learning for computer aided diagnosis

Dijia Wu, Jinbo Bi, K. Boyer
2009 2009 IEEE Conference on Computer Vision and Pattern Recognition  
To address all these problems, we propose a novel learning formulation to combine cascade classification and multiple instance learning (MIL) in a unified min-max framework, leading to a joint optimization  ...  data between negative and positive classes; stringent real-time requirement of online execution; multiple positive candidates generated for the same malignant structure that are highly correlated and  ...  bag as positive, which is different from a standard MIL problem where instances in a positive bag can be unlabeled or even negative.  ... 
doi:10.1109/cvprw.2009.5206778 fatcat:lphpxadi2zcwbgezxm7jxk5rj4

A min-max framework of cascaded classifier with multiple instance learning for computer aided diagnosis

Dijia Wu, Jinbo Bi, Kim Boyer
2009 2009 IEEE Conference on Computer Vision and Pattern Recognition  
To address all these problems, we propose a novel learning formulation to combine cascade classification and multiple instance learning (MIL) in a unified min-max framework, leading to a joint optimization  ...  data between negative and positive classes; stringent real-time requirement of online execution; multiple positive candidates generated for the same malignant structure that are highly correlated and  ...  bag as positive, which is different from a standard MIL problem where instances in a positive bag can be unlabeled or even negative.  ... 
doi:10.1109/cvpr.2009.5206778 dblp:conf/cvpr/WuBB09 fatcat:gbamnl6f3zfqvjmofxlb6bsscy
« Previous Showing results 1 — 15 out of 1,854 results