Filters








13,286 Hits in 5.0 sec

Adaptive Hierarchical Similarity Metric Learning with Noisy Labels [article]

Jiexi Yan, Lei Luo, Cheng Deng, Heng Huang
2021 arXiv   pre-print
However, most existing deep metric learning methods with binary similarity are sensitive to noisy labels, which are widely present in real-world data.  ...  In this paper, we propose an Adaptive Hierarchical Similarity Metric Learning method. It considers two noise-insensitive information, i.e., class-wise divergence and sample-wise consistency.  ...  CONCLUSION In this paper, we proposed a novel Adaptive Hierarchical Similarity Metric Learning approach combating noisy labels.  ... 
arXiv:2111.00006v1 fatcat:nx6hsswcijgd3ba6q5umrip7lu

Dual-Refinement: Joint Label and Feature Refinement for Unsupervised Domain Adaptive Person Re-Identification [article]

Yongxing Dai, Jun Liu, Yan Bai, Zekun Tong, Ling-Yu Duan
2021 arXiv   pre-print
Our Dual-Refinement method reduces the influence of noisy labels and refines the learned features within the alternative training process.  ...  Unsupervised domain adaptive (UDA) person re-identification (re-ID) is a challenging task due to the missing of labels for the target domain data.  ...  Learning with Noisy Labels Existing works on learning with noisy labels can be categorized into four main groups.  ... 
arXiv:2012.13689v2 fatcat:eowgknmyuzep7axqgontxoesl4

Multi-source Hierarchical Prediction Consolidation [article]

Chenwei Zhang, Sihong Xie, Yaliang Li, Jing Gao, Wei Fan, Philip S. Yu
2016 arXiv   pre-print
We propose a novel multi-source hierarchical prediction consolidation method to effectively exploits the complicated hierarchical label structures to resolve the noisy and conflicting information that  ...  Although state-of-the-art aggregation methods have been proposed to handle label spaces with flat structures, as the label space is becoming more and more complicated, aggregation under a label hierarchical  ...  The multi-source label space contains multi-source hierarchical label predictions with noisy and conflicting labels.  ... 
arXiv:1608.03344v1 fatcat:psiavl22orfrpnais36w3iebga

Unsupervised Semantic Scene Labeling for Streaming Data

Maggie Wigness, John G. Rogers
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
We introduce an unsupervised semantic scene labeling approach that continuously learns and adapts semantic models discovered within a data stream.  ...  algorithms given similar numbers of label outputs.  ...  During evaluation, some metrics are computed across the entire hierarchical output of GBH and S-GBH, but we focus on comparisons using the hierarchical segmentation level that produces a similar number  ... 
doi:10.1109/cvpr.2017.626 dblp:conf/cvpr/WignessR17 fatcat:u3nhr7k52jf65kq5zab7dy2ha4

Learning Category Correlations for Multi-label Image Recognition with Graph Networks [article]

Qing Li, Xiaojiang Peng, Yu Qiao, Qiang Peng
2019 arXiv   pre-print
with an Adaptive label correlation graph to model label dependencies.  ...  Specifically, we introduce a plug-and-play Label Graph (LG) module to learn label correlations with word embeddings, and then utilize traditional GCN to map this graph into label-dependent object classifiers  ...  (Wu et al. 2015) propose an approach named weakly semi-supervised deep learning for multi-label image annotation, which uses a triplet loss function to draw images with similar label sets.  ... 
arXiv:1909.13005v1 fatcat:onvqczvlb5f3nmvtzdj7hyavby

Uncertainty-aware Clustering for Unsupervised Domain Adaptive Object Re-identification [article]

Pengfei Wang, Changxing Ding, Wentao Tan, Mingming Gong, Kui Jia, Dacheng Tao
2021 arXiv   pre-print
Unsupervised Domain Adaptive (UDA) object re-identification (Re-ID) aims at adapting a model trained on a labeled source domain to an unlabeled target domain.  ...  Second, an uncertainty-aware collaborative instance selection method is introduced to select images with reliable labels for model training.  ...  Related Works We review the literature in three parts: 1) unsupervised domain adaptive (UDA) object Re-ID, 2) contrastive learning, and 3) deep learning with noisy labels.  ... 
arXiv:2108.09682v1 fatcat:7kxdkzzf3fevfdcxjmt6eox3nu

A Review article on Semi- Supervised Clustering Framework for High Dimensional Data

M. Pavithra, R. M. S. Parvathi
2019 International Journal of Scientific Research in Computer Science Engineering and Information Technology  
supervised clustering, with minimum labeled data, self-organizing based on neural networks.  ...  Clustering algorithms are based on active learning, with ensemble clustering-means algorithm, data streams with flock, fuzzy clustering for shape annotations, Incremental semi supervised clustering, Weakly  ...  In general, the metric learning used in the distance based method, which is equivalent to learning an adaptive weight for each dimension, is either based on iterative algorithms, such as gradient descent  ... 
doi:10.32628/cseit195410 fatcat:xl37f2eb6bagjfwdscc7fwi2p4

Incremental Figure-Ground Segmentation Using Localized Adaptive Metrics in LVQ [chapter]

Alexander Denecke, Heiko Wersing, Jochen J. Steil, Edgar Körner
2009 Lecture Notes in Computer Science  
In presence of local adaptive metrics and supervised noisy information we use a parallel evaluation scheme combined with a local utility function to organize a learning vector quantization (LVQ) network  ...  with an adaptive number of prototypes and verify the capabilities on a real world figure-ground segmentation task.  ...  In presence of local adaptive metrics and supervised noisy information we use a parallel evaluation scheme combined with a local utility function to organize a learning vector quantization with an adaptive  ... 
doi:10.1007/978-3-642-02397-2_6 fatcat:s5vgqhto3varhozdmisegqbjoi

Autonomous Learning of Representations

Oliver Walter, Reinhold Haeb-Umbach, Bassam Mokbel, Benjamin Paassen, Barbara Hammer
2015 Künstliche Intelligenz  
The goal of this contribution is to give an overview about different principles of autonomous feature learning, and to exemplify two principles based on two recent examples: autonomous metric learning  ...  Besides the core learning algorithm itself, one major question in machine learning is how to best encode given training data such that the learning technology can efficiently learn based thereon and generalize  ...  A combination with metric learning and clustering of similar words with different pronunciations or spelling errors, could improve the learning result.  ... 
doi:10.1007/s13218-015-0372-1 fatcat:gnttnjqv7basviulnwx573jxue

Prohibited Item Detection via Risk Graph Structure Learning

Yugang Ji, Guanyi Chu, Xiao Wang, Chuan Shi, Jianan Zhao, Junping Du
2022 Proceedings of the ACM Web Conference 2022  
RGSL first introduces structure learning into large-scale risk graphs, to reduce noisy connections and add similar pairs.  ...  It then designs the pairwise training mechanism, which transforms the detection process as a metric learning from candidates to their similar prohibited items.  ...  A basic idea is to introduce metric learning to learn the distance between labeled items.  ... 
doi:10.1145/3485447.3512190 fatcat:gk6jk7hhojgmle2bekwroub6by

HiURE: Hierarchical Exemplar Contrastive Learning for Unsupervised Relation Extraction [article]

Shuliang Liu, Xuming Hu, Chenwei Zhang, Shu`ang Li, Lijie Wen, Philip S. Yu
2022 arXiv   pre-print
instance-wise contrastive learning which unreasonably pushes apart those sentence pairs that are semantically similar.  ...  To overcome these defects, we propose a novel contrastive learning framework named HiURE, which has the capability to derive hierarchical signals from relational feature space using cross hierarchy attention  ...  Hierarchical Exemplar Contrastive Learning In order to adaptively generate more positive samples other than sentences themselves to introduce more similarity information in contrastive learning, we design  ... 
arXiv:2205.02225v2 fatcat:klqqsnruurcaxfnsv5ujb43mvu

A Literature Review of Gene Function Prediction by Modeling Gene Ontology

Yingwen Zhao, Jun Wang, Jian Chen, Xiangliang Zhang, Maozu Guo, Guoxian Yu
2020 Frontiers in Genetics  
quantifying semantic similarities.  ...  To bridge this gap, we review the existing methods with an emphasis on recent solutions.  ...  the similarity between genes to identify noisy annotations.  ... 
doi:10.3389/fgene.2020.00400 pmid:32391061 pmcid:PMC7193026 fatcat:u3jc3ieejzebdfxlbfhrvdbvp4

Meta learning to classify intent and slot labels with noisy few shot examples [article]

Shang-Wen Li, Jason Krone, Shuyan Dong, Yi Zhang, Yaser Al-onaizan
2020 arXiv   pre-print
(IC) and slot labeling (SL).  ...  To improve the performance of SLU models on tasks with noisy and low training resources, we propose a new SLU benchmarking task: few-shot robust SLU, where SLU comprises two core problems, intent classification  ...  Adaptation of f θ to new domains is achieved with a few labeled examples provided by the support set, whereas model performance is evaluated by averaging metrics measured in each episode's query set.  ... 
arXiv:2012.07516v1 fatcat:hrsgnmpf4fevvh2kdefvircymi

Weakly Supervised Learning with Side Information for Noisy Labeled Images [article]

Lele Cheng, Xiangzeng Zhou, Liming Zhao, Dangwei Li, Hong Shang, Yun Zheng, Pan Pan, Yinghui Xu
2020 arXiv   pre-print
In this paper, we present an efficient weakly supervised learning by using a Side Information Network (SINet), which aims to effectively carry out a large scale classification with severely noisy labels  ...  In many real-world datasets, like WebVision, the performance of DNN based classifier is often limited by the noisy labeled data.  ...  Self-learning pseudo-labels has been studied in many scenarios to deal with noisy labels. Reed et al. [17] propose to jointly train model with both noisy labels and pseudo-labels.  ... 
arXiv:2008.11586v2 fatcat:336pp4msefeatjnt6bxd2k73ry

Creating a Cluster Hierarchy under Constraints of a Partially Known Hierarchy [chapter]

Korinna Bade, Andreas Nürnberger
2008 Proceedings of the 2008 SIAM International Conference on Data Mining  
The approaches cover the two major fields of constraint based clustering, i.e. instance and metric based constraint integration. Our objects of interest are text documents.  ...  We introduce the concept of hierarchical constraints and continue by presenting and evaluating two approaches using them.  ...  While the metric is usually learned in advance using only the given constraints, the approach in [5] adapts the distance metric during clustering.  ... 
doi:10.1137/1.9781611972788.2 dblp:conf/sdm/BadN08 fatcat:2efx2etdj5ex5hpzgqcqfaqjye
« Previous Showing results 1 — 15 out of 13,286 results