482 Hits in 4.2 sec

Convex and Scalable Weakly Labeled SVMs [article]

Yu-Feng Li, Ivor W. Tsang, James T. Kwok, Zhi-Hua Zhou
2013 arXiv   pre-print
Therefore, it can suffer from poor scalability and may also get stuck in local minimum. In this paper, we focus on SVMs and propose the WellSVM via a novel label generation strategy.  ...  Moreover, the WellSVM can be solved via a sequence of SVM subproblems that are much more scalable than previous convex SDP relaxations.  ...  ., 2009c,a) , we propose a convex weakly labeled SVM (denoted WellSVM (WEakly LabeLed SVM)) via a novel "label generation" strategy.  ... 
arXiv:1303.1271v5 fatcat:3xqbttwx5rffjmifidls5tapfi

A Convex Relaxation for Weakly Supervised Classifiers [article]

Armand Joulin, Francis Bach (INRIA - Ecole Normale Superieure)
2012 arXiv   pre-print
Inferring the labels and learning the parameters of the model is usually done jointly through a block-coordinate descent algorithm such as expectation-maximization (EM), which may lead to local minima.  ...  This paper introduces a general multi-class approach to weakly supervised classification.  ...  This paper was partially supported by the European Research Council (SIERRA and VIDEOWORLD projects).  ... 
arXiv:1206.6413v1 fatcat:jroc2kovr5e7zdueu6ayvjitfa

Multi-instance Methods for Partially Supervised Image Segmentation [chapter]

Andreas Müller, Sven Behnke
2012 Lecture Notes in Computer Science  
Using these candidate segments, we solve the multi-instance, multi-class problem using multi-instance kernels with an SVM.  ...  This computationally advantageous approach, which requires only convex optimization, yields encouraging results on the challenging problem of partially supervised image segmentation.  ...  [10] , so we use the MI-kernel for better scalability. Li and Sminchisescu [17] compute likelihood ratios for instances, giving a new convex formulation of the multi-instance problem.  ... 
doi:10.1007/978-3-642-28258-4_12 fatcat:a5knctennjbahcnqxbld7e33s4

Feature and Region Selection for Visual Learning

Ji Zhao, Liantao Wang, Ricardo Cabral, Fernando De la Torre
2016 IEEE Transactions on Image Processing  
in images and spatio-temporal regions in videos in a unified way; (3) the feature selection problem is convex, and both problems can be solved using a scalable reduced gradient method; (4) we point out  ...  strong connections with multiple kernel learning and multiple instance learning approaches.  ...  Problem (7) is convex, and we propose a scalable optimization strategy in Section III-D. C.  ... 
doi:10.1109/tip.2016.2514503 pmid:26742135 fatcat:6v7zgdldzveeblnd5zs6bgsupm

Strong supervision from weak annotation: Interactive training of deformable part models

Steve Branson, Pietro Perona, Serge Belongie
2011 2011 International Conference on Computer Vision  
This framework is scalable to large datasets and complex image models and is shown to have excellent theoretical and practical properties in terms of train time, optimality guarantees, and bounds on the  ...  The system interleaves interactive labeling (where the current model is used to semiautomate the labeling of a new example) and online learning (where a newly labeled example is used to update the current  ...  The authors thank Boris Babenko, Kristin Branson, and Peter Welinder for helpful discussions and feedback.  ... 
doi:10.1109/iccv.2011.6126450 dblp:conf/iccv/BransonPB11 fatcat:xyku6nqekzgbhh5xp4khti5xei

Audio Event and Scene Recognition: A Unified Approach using Strongly and Weakly Labeled Data [article]

Anurag Kumar, Bhiksha Raj
2017 arXiv   pre-print
In this paper we propose a novel learning framework called Supervised and Weakly Supervised Learning where the goal is to learn simultaneously from weakly and strongly labeled data.  ...  In weakly supervised learning only data is weakly labeled which prevents one from directly applying supervised learning methods.  ...  [17] used SVM (miSVM) [22] and neural network (BPMIL) [23] based MIL methods to show that AED with weak labels can be successfully done.  ... 
arXiv:1611.04871v3 fatcat:jclmmvkhzrhyfgjygjpxxafw6a

Convex Learning with Invariances

Choon Hui Teo, Amir Globerson, Sam T. Roweis, Alexander J. Smola
2007 Neural Information Processing Systems  
We provide a convex formulation which can deal with arbitrary loss functions and arbitrary losses.  ...  Acknowledgements: We thank Carlos Guestin and Bob Williamson for fruitful discussions. Part of the work was done when CHT was visiting NEC Labs America.  ...  original samples (STD-SVM), SVM trained on original and virtual samples (VIR-SVM), and our convex invariance-loss method (Invar-SVM).  ... 
dblp:conf/nips/TeoGRS07 fatcat:ujioevnjcnby5mjpyk2mlms37a

Semi-supervised Learning with Weakly-Related Unlabeled Data: Towards Better Text Categorization

Liu Yang, Rong Jin, Rahul Sukthankar
2008 Neural Information Processing Systems  
To this end, SSLW estimates the optimal wordcorrelation matrix that is consistent with both the co-occurrence information derived from the weakly-related unlabeled documents and the labeled documents.  ...  We show that SSLW results in a significant improvement in categorization accuracy, equipped with a small training set and an unlabeled resource that is weakly related to the test domain.  ...  Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of NSF and NIH.  ... 
dblp:conf/nips/YangJS08 fatcat:z5crr4ryovatbiivzvbhrmvu24

Latent SVM for Object Localization in Weakly Labeled Videos

Mrigank Rochan, Yang Wang
2015 2015 12th Conference on Computer and Robot Vision  
Such videos are weakly labeled. Given weakly labeled video with video-level object class tags, our goal is to learn a model that can be used to localize the objects in other videos with such tags.  ...  We define a latent SVM based learning framework to tackle this problem. We demonstrate the effectiveness of our method on a dataset composed of videos collected from YouTube.  ...  Therefore, our method is scalable and can be used to exploit the huge amount of weakly labeled video data available on Internet to address various challenges in video understanding.  ... 
doi:10.1109/crv.2015.33 dblp:conf/crv/RochanW15 fatcat:a3qqvidsmvcnbhzhzjci3zu5j4

Sequential Alternating Proximal Method for Scalable Sparse Structural SVMs

P. Balamurugan, Shirish Shevade, T. Ravindra Babu
2012 2012 IEEE 12th International Conference on Data Mining  
SVMs are used on very large datasets.  ...  Though L1-regularized structural SVMs have been studied in the past, the use of elastic net regularizer for structural SVMs has not been explored yet.  ...  STRUCTURAL SVMS Structural SVMs [15] [16] learn from training data {(x i , y i )} n i=1 of structured inputs and outputs, by finding the solution to a convex Quadratic Program (QP), which is (OP1): min  ... 
doi:10.1109/icdm.2012.81 dblp:conf/icdm/PSB12 fatcat:bfd5pv4amvantgu2j7veoj7idm

Weakly supervised sparse coding with geometric consistency pooling

Liujuan Cao, Rongrong Ji, Yue Gao, Yi Yang, Qi Tian
2012 2012 IEEE Conference on Computer Vision and Pattern Recognition  
First, we introduce a Weakly supervised Sparse Coding (WSC) to exploit the Classemes-based attribute labeling to refine the descriptor coding procedure.  ...  Therefore, our approach enables potential applications like scalable visual search.  ...  Bi-Convex Optimization Learning both sparse codes V (WSC + GCP) and dictionary U are not convex simultaneously. Following [2] [15], we learn V and U alternatively.  ... 
doi:10.1109/cvpr.2012.6248102 dblp:conf/cvpr/CaoJGYT12 fatcat:tpgmreckkjfmhdzmzc3fsvqhtm

Multi-label Discriminative Weakly-Supervised Human Activity Recognition and Localization [chapter]

Ehsan Adeli Mosabbeb, Ricardo Cabral, Fernando De la Torre, Mahmood Fathy
2015 Lecture Notes in Computer Science  
Recently, weakly-supervised learning (WSL) approaches were able to learn discriminative classifiers while localizing the action in space and/or time using weak labels.  ...  To date, most approaches to video rely on fully supervised settings that require time consuming and error prone manual labeling.  ...  [8] trained a latent SVM with a number of labeled and fully annotated videos, but each video is assigned a single label.  ... 
doi:10.1007/978-3-319-16814-2_16 fatcat:zb3ydfn3r5doxopbbgeyeuvpce

Weakly supervised object detection with convex clustering

Hakan Bilen, Marco Pedersoli, Tinne Tuytelaars
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Weakly supervised object detection, is a challenging task, where the training procedure involves learning at the same time both, the model appearance and the object location in each image.  ...  However, as learning appearance and localization are two interconnected tasks, the optimization is not convex and the procedure can easily get stuck in a poor local minimum, i.e. the algorithm "misses"  ...  Finally, we want our method to be scalable.  ... 
doi:10.1109/cvpr.2015.7298711 dblp:conf/cvpr/BilenPT15 fatcat:f3keimbbjfcdvlzdei2arcun4i

Weakly supervised discriminative localization and classification: a joint learning process

Minh Hoai Nguyen, Lorenzo Torresani, Fernando de la Torre, Carsten Rother
2009 2009 IEEE 12th International Conference on Computer Vision  
the data needed, and completing and reviewing the collection of information.  ...  Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining  ...  Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors  ... 
doi:10.1109/iccv.2009.5459426 dblp:conf/iccv/NguyenTTR09 fatcat:z6abgouwybfzdlxrbt6deqemey

Weakly Supervised Visual Dictionary Learning by Harnessing Image Attributes

Yue Gao, Rongrong Ji, Wei Liu, Qionghai Dai, Gang Hua
2014 IEEE Transactions on Image Processing  
However, such labels are typically too expensive to acquire, which restricts the scalability of supervised dictionary learning approaches.  ...  In this paper, we propose to leverage image attributes to weakly supervise the dictionary learning procedure without requiring any actual labels.  ...  Sift) + stochastic SVM, and finally (4) BoF (Dense Sift) + stochastic SVM.  ... 
doi:10.1109/tip.2014.2364536 pmid:25361504 fatcat:vfuulvkvbvd6pjpn3libbzria4
« Previous Showing results 1 — 15 out of 482 results