Filters








6,856 Hits in 6.5 sec

Curriculum Learning of Visual Attribute Clusters for Multi-Task Classification [article]

Nikolaos Sarafianos, Theodore Giannakopoulos, Christophoros Nikou, Ioannis A. Kakadiaris
2018 arXiv   pre-print
In this paper, we introduce a novel method to combine the advantages of both multi-task and curriculum learning in a visual attribute classification framework.  ...  The clusters of tasks are learned in a curriculum learning setup by transferring knowledge between clusters.  ...  All statements of fact, opinion or conclusions contained herein are those of the authors and should not be construed as representing the official views or policies of the sponsors.  ... 
arXiv:1709.06664v3 fatcat:b3i7y7x63nb7bjbaxedviwwmry

Curriculum Learning for Multi-Task Classification of Visual Attributes [article]

Nikolaos Sarafianos, Theodore Giannakopoulos, Christophoros Nikou, Ioannis A. Kakadiaris
2017 arXiv   pre-print
In this paper, we introduce a novel method to combine the advantages of both multi-task and curriculum learning in a visual attribute classification framework.  ...  The learning process within each group though, is performed in a multi-task classification setup.  ...  All statements of fact, opinion or conclusions contained herein are those of the authors and should not be construed as representing the official views or policies of the sponsors.  ... 
arXiv:1708.08728v1 fatcat:b6zugyrmhzcpphuvmok5tkyfyy

Curriculum Learning for Multi-task Classification of Visual Attributes

Nikolaos Sarafianos, Theodore Giannakopoulos, Christophoros Nikou, Ioannis A. Kakadiaris
2017 2017 IEEE International Conference on Computer Vision Workshops (ICCVW)  
In this paper, we introduce a novel method to combine the advantages of both multi-task and curriculum learning in a visual attribute classification framework.  ...  The two groups of tasks are learned in a curriculum learning setup by transferring the acquired knowledge from the strongly to the weakly correlated.  ...  All statements of fact, opinion or conclusions contained herein are those of the authors and should not be construed as representing the official views or policies of the sponsors.  ... 
doi:10.1109/iccvw.2017.306 dblp:conf/iccvw/SarafianosGNK17 fatcat:svb54a2sdvgnplmbrifcys6qqq

Attention-Guided Curriculum Learning for Weakly Supervised Classification and Localization of Thoracic Diseases on Chest Radiographs [chapter]

Yuxing Tang, Xiaosong Wang, Adam P. Harrison, Le Lu, Jing Xiao, Ronald M. Summers
2018 Lecture Notes in Computer Science  
A convolutional neural network (CNN) based attention-guided curriculum learning (AGCL) framework is presented, which leverages the severity-level attributes mined from radiology reports.  ...  AGCL achieves over 5.7% (averaged over 14 diseases) increase in classification AUC and 7%/11% increases in Recall/Precision for the localization task compared to the state of the art.  ...  The authors thank NVIDIA for GPU donation.  ... 
doi:10.1007/978-3-030-00919-9_29 fatcat:ta4ehmmajndepcfm2qq3j2jwi4

Pedestrian Attribute Recognition: A Survey [article]

Xiao Wang, Shaofei Zheng, Rui Yang, Bin Luo, Jin Tang
2019 arXiv   pre-print
Thirdly, we analyse the concept of multi-task learning and multi-label learning, and also explain the relations between these two learning algorithms and pedestrian attribute recognition.  ...  Fourthly, we analyse popular solutions for this task, such as attributes group, part-based, etc.  ...  Once the total dependencies for all the clusters formed, the curriculum learning process can be started in a descending order.  ... 
arXiv:1901.07474v1 fatcat:h5krexsotbecvlwi2w4uw2y3ay

Label-similarity Curriculum Learning [article]

Urun Dogan, Aniket Anand Deshmukh, Marcin Machura, Christian Igel
2020 arXiv   pre-print
The proposed label-similarity curriculum learning (LCL) approach was empirically evaluated using several popular deep learning architectures for image classification tasks applied to five datasets including  ...  We propose a novel curriculum learning approach for image classification that adapts the loss function by changing the label representation.  ...  Distinguishing between these clusters of classes is an easier task than distinguishing between classes within one cluster. Classification Performance.  ... 
arXiv:1911.06902v2 fatcat:bg62yxchkbhszdwppvxaa6rfpq

An Automatic Classification and Clustering Algorithm for Online Learning Goals Based on Cognitive Thinking

Ying Wang, Weifeng Jiang
2018 International Journal of Emerging Technologies in Learning (iJET)  
The results showed that the proposed algorithm for automatic classification and clustering of online learning targets had a good application effect in the learning community.  ...  To improve the learning effect of online learning, an online learning target automatic classification and clustering analysis algorithm based on cognitive thinking was proposed.  ...  The goal of cluster analysis is to collect data on a similar basis for classification.  ... 
doi:10.3991/ijet.v13i11.9587 fatcat:si6oztgy2fh5ljxg4g7lhm7c3u

Curriculum Learning: A Survey [article]

Petru Soviany, Radu Tudor Ionescu, Paolo Rota, Nicu Sebe
2022 arXiv   pre-print
We construct a multi-perspective taxonomy of curriculum learning approaches by hand, considering various classification criteria.  ...  Curriculum learning strategies have been successfully employed in all areas of machine learning, in a wide range of tasks.  ...  Acknowledgements The authors would like to thank the reviewers for their useful feedback.  ... 
arXiv:2101.10382v3 fatcat:doognr7ggfaalg7kd2i3n7s3jy

Data Mining Approach towards Students Behavior Assessment Methods for Higher Studies

Sharad Gangele, Kirti Soni, Sunil Patil
2018 International Journal of Computer Applications  
on learning method.  ...  The proposed framework can be applied to extract valuable data that shows all characteristic of student behavior by clustering and subdivision of the student behavior large data set.  ...  To make classification, the algorithm uses the criteria of information gain. The more the information gain of an attribute, the more is the split possibility of that attribute.  ... 
doi:10.5120/ijca2018918099 fatcat:56vlnxokt5hohdvrqpiyw3zyuy

An Educational Data Mining Model for Predicting Student Performance in Programming Course

A. F.ElGamal
2013 International Journal of Computer Applications  
The proposed model includes three phases; data preprocessing, attribute selection and rule extraction algorithm.  ...  This paper presents an educational data mining model for predicting student performance in programming courses. Identifying variables that predict student programming performance may help educators.  ...  The general tasks of classification, regression, clustering, or deviation analysis have a large number of solutions such as neural networks, decision tree learners, rule learners or Bayesian networks [  ... 
doi:10.5120/12160-8163 fatcat:kvrnxplqfjejbhpudcdj5ckvou

Dynamic Curriculum Learning for Imbalanced Data Classification

Yiru Wang, Weihao Gan, Jie Yang, Wei Wu, Junjie Yan
2019 2019 IEEE/CVF International Conference on Computer Vision (ICCV)  
Human attribute analysis is a challenging task in the field of computer vision. One of the significant difficulties is brought from largely imbalance-distributed data.  ...  ; (2) loss scheduler which controls the learning importance between classification and metric learning loss.  ...  Method We propose a Dynamic Curriculum Learning (DCL) framework for imbalanced data classification problem, consisting of two-level curriculum schedulers.  ... 
doi:10.1109/iccv.2019.00512 dblp:conf/iccv/WangGYWY19 fatcat:kzthfmteorerdkbsg4h2ogtl24

Dynamic Curriculum Learning for Imbalanced Data Classification [article]

Yiru Wang, Weihao Gan, Jie Yang, Wei Wu, Junjie Yan
2019 arXiv   pre-print
Human attribute analysis is a challenging task in the field of computer vision, since the data is largely imbalance-distributed.  ...  ; (2) loss scheduler controls the learning importance between classification and metric learning loss.  ...  Method We propose a Dynamic Curriculum Learning (DCL) framework for imbalanced data classification problem, consisting of two-level curriculum schedulers.  ... 
arXiv:1901.06783v2 fatcat:7mdcol5jjvhtzdnzf5ls7fvwdm

Visual Word2Vec (vis-w2v): Learning Visually Grounded Word Embeddings Using Abstract Scenes [article]

Satwik Kottur, Ramakrishna Vedantam, José M. F. Moura, Devi Parikh
2016 arXiv   pre-print
We propose a model to learn visually grounded word embeddings (vis-w2v) to capture visual notions of semantic relatedness.  ...  We show improvements over text-only word embeddings (word2vec) on three tasks: common-sense assertion classification, visual paraphrasing and text-based image retrieval.  ...  Other works use visual and textual attributes (e.g. vegetable is an attribute for potato) to improve distributional models of word meaning [38, 39] .  ... 
arXiv:1511.07067v2 fatcat:5qplkhak35du7dt7nbmqo2em6e

Learning Words by Drawing Images

Didac Suris, Adria Recasens, David Bau, David Harwath, James Glass, Antonio Torralba
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
We propose a framework for learning through drawing. Our goal is to learn the correspondence between spoken words and abstract visual attributes, from a dataset of spoken descriptions of images.  ...  We find that training that takes advantage of GANgenerated edited examples results in improvements in the model's ability to learn attributes compared to previous results.  ...  Other works have also demonstrated the utility of audio-visual features for supervised classification tasks [4, 6] , or predicting the sound made by an object [36, 37] .  ... 
doi:10.1109/cvpr.2019.00213 dblp:conf/cvpr/SurisRBHG019 fatcat:xz3cukvvnzgzji3htbjjdmvfuq

Fine-Grain Few-Shot Vision via Domain Knowledge as Hyperspherical Priors [article]

Bijan Haney, Alexander Lavin
2020 arXiv   pre-print
Prototypical networks have been shown to perform well at few-shot learning tasks in computer vision.  ...  We describe how to construct a hypersphere of prototypes that embed a-priori domain information, and demonstrate the effectiveness of the approach on challenging benchmark datasets for fine-grain classification  ...  Effective methods use CNNs with innovative learning approaches such as multi-scale [16] , transfer [5] , and curriculum [1] learning. Figure 1 .  ... 
arXiv:2005.11450v1 fatcat:aek4seqrrvejdi42axdthft6o4
« Previous Showing results 1 — 15 out of 6,856 results