Filters








4,750 Hits in 4.7 sec

Weakly Supervised Learning by a Confusion Matrix of Contexts [chapter]

William Wu
2019 Lecture Notes in Computer Science  
Context consideration can help provide more background and related information for weakly supervised learning.  ...  for more informative patterns in order to improve weakly supervised learning from limited labeled samples for unlabeled data.  ...  It is hoped that this is only the beginning, and other factors can later be introduced to this multi-contextual model to expand the scope of data exploration and knowledge discovery for weakly supervised  ... 
doi:10.1007/978-3-030-26142-9_6 fatcat:27tyndeqhjbcrksv2bfnfeybte

Classifying Unstructured Clinical Notes via Automatic Weak Supervision [article]

Chufan Gao, Mononito Goswami, Jieshi Chen, Artur Dubrawski
2022 arXiv   pre-print
In this work, we introduce a general weakly-supervised text classification framework that learns from class-label descriptions only, without the need to use any human-labeled documents.  ...  We demonstrate the efficacy and flexibility of our method by comparing it to state-of-the-art weak text classifiers across four real-world text classification datasets, in addition to assigning ICD codes  ...  fully-supervised (Devlin et al., 2019 ) models on four real-world text classification problems.  ... 
arXiv:2206.12088v2 fatcat:tu4xqyaalffw7mzdvv7ulwolke

YouCat: Weakly Supervised Youtube Video Categorization System from Meta Data & User Comments using WordNet & Wikipedia

Subhabrata Mukherjee, Pushpak Bhattacharyya
2012 International Conference on Computational Linguistics  
In this paper, we propose a weakly supervised system, YouCat, for categorizing Youtube videos into different genres like Comedy, Horror, Romance, Sports and Technology The system takes a Youtube video  ...  Experimentations have been performed on real world Youtube videos where YouCat achieves an F-score of 80.9%, without using any labeled training set, compared to the supervised, multiclass SVM F-score of  ...  WordNet and Wikipedia are used as world knowledge sources for expanding the video descriptor since the uploader provided text is frequently very short, as are the user comments.  ... 
dblp:conf/coling/MukherjeeB12a fatcat:axfz373xbzgtplpn6gowmdousu

Long-Term Memory Networks for Question Answering [article]

Fenglong Ma, Radha Chitta, Saurabh Kataria, Jing Zhou, Palghat Ramesh, Tong Sun, Jing Gao
2017 arXiv   pre-print
The LTMN model can be trained end-to-end using back-propagation and requires minimal supervision.  ...  Several deep neural network architectures have been developed recently, which employ memory and inference components to memorize and reason over text information, and generate answers to questions.  ...  Traditional question answering approaches can be categorized as: (i) IR-based question answering [Paşca, 2003] where the question is formulated as a search query, and a short text segment is found on  ... 
arXiv:1707.01961v1 fatcat:6jiqnvogojd2jabgfijy4js7fe

Machine Learning with World Knowledge: The Position and Survey [article]

Yangqiu Song, Dan Roth
2017 arXiv   pre-print
We start from the comparison of world knowledge with domain-specific knowledge, and then introduce three key problems in using world knowledge in learning processes, i.e., explicit and implicit feature  ...  In this paper, we will discuss how to use the existing general-purpose world knowledge to enhance machine learning processes, by enriching the features or reducing the labeling work.  ...  We categorize the paradigms into ways related to world knowledge features and ways related to world knowledge supervision.  ... 
arXiv:1705.02908v1 fatcat:t4fypa6h3vampcp64eosvppsfe

Adaptive Named Entity Recognition Using Distant Supervision for Contemporary Written Texts

Juae Kim, Yejin Kim, Sangwoo Kang, Jungyun Seo
2021 IEEE Access  
Named entity recognition (NER) is the process of categorizing named entities in a given text that suffers from the lack of labeled corpora, which is a long-standing issue.  ...  We evaluated the proposed methods with bidirectional long short-term memory conditional random-field architecture using static and contextualized embedding methods.  ...  Using weakly supervised data is advantageous in that human labeling is not required as long as there is a knowledge source full of unlabeled data.  ... 
doi:10.1109/access.2021.3067315 fatcat:3pyfiggtqbhhvonam4g2sxhdii

Complex Relation Extraction: Challenges and Opportunities [article]

Haiyun Jiang, Qiaoben Bao, Qiao Cheng, Deqing Yang, Li Wang, Yanghua Xiao
2020 arXiv   pre-print
Relation extraction aims to identify the target relations of entities in texts. Relation extraction is very important for knowledge base construction and text understanding.  ...  Traditional binary relation extraction, including supervised, semi-supervised and distant supervised ones, has been extensively studied and significant results are achieved.  ...  In general, BiRE can be categorized into: supervised, semi-supervised, distant supervised paradigms. Supervised BiRE Description.  ... 
arXiv:2012.04821v1 fatcat:6fnhjnpwmrabhhrfit3x63tfnm

Sprinkling Topics for Weakly Supervised Text Classification

Swapnil Hingmire, Sutanu Chakraborti
2014 Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)  
We evaluate this approach to improve performance of text classification on three real world datasets.  ...  In this paper, we propose a weakly supervised algorithm in which supervision comes in the form of labeling of Latent Dirichlet Allocation (LDA) topics.  ...  These algorithms can be broadly categorized into three categories depending on how supervision is provided.  ... 
doi:10.3115/v1/p14-2010 dblp:conf/acl/HingmireC14 fatcat:jvvvxvhb7vbj5lgfaw3hcvbynu

Analysis of User's Behavioral Pattern for Various Brands Using Machine Learning

Farah Jamal,, Ms. Kavita Agrawal
2019 International Journal of Research in Advent Technology  
The main aim is to classify the text as positive, negative or neutral. Various sentiment classifier are used to analyze the data.  ...  Here, Naïve Bayes classifier is used for classification, Long-Short term memory for predicting the outcome and decision tree for precision.  ...  We will be using Naïve Bayes, LSTM and Decision Tree. Sentiment Analysis approaches can be categorized as: -Machine learning, lexicon-based and hybrid.  ... 
doi:10.32622/ijrat.752019316 fatcat:tp64ygt4sjfefndpkj5w3ktuvy

Why Biomedical Relation Extraction Is an Open Issue?

Farzane Ghamami, Mohammadreza Keyvanpour
2018 Innovative Computing Information and Control Express Letters, Part B: Applications  
One of the techniques for using this information includes relation extractions. Biomedical relation extraction methods gather valuable, reliable, and affordable information from free-form texts.  ...  Articles on biomedical area are available in free-form texts in libraries and databases. Many articles published on biomedicine field contain up-to-date information.  ...  Finally, (×) mark means the challenge exists as an × × × × × × × × Self Supervised × ↑ × × ↑ ↑ ↑ × ↑ × × ↑ Joint Models × ↑ ↑ ↑ ↑ × ↑ × × ↑ ↑ Knowledge Based × × × × × × × ↑ × × Weakly Supervised × × ×  ... 
doi:10.24507/icicelb.09.08.747 fatcat:nwfb5nqhfvaxdjydfkt3n47iou

Learning Deep Features for Discriminative Localization

Bolei Zhou, Aditya Khosla, Agata Lapedriza, Aude Oliva, Antonio Torralba
2016 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
apparent simplicity of global average pooling, we are able to achieve 37.1% top-5 error for object localization on ILSVRC 2014, which is remarkably close to the 34.2% top-5 error achieved by a fully supervised  ...  Weakly supervised text detector: We train a weakly supervised text detector using 350 Google StreetView images containing text from the SVT dataset [25] as the positive set and randomly sampled images  ...  Figure 11 . 11 Learning a weakly supervised text detector. The text is accurately detected on the image even though our network is not trained with text or any bounding box annotations.  ... 
doi:10.1109/cvpr.2016.319 dblp:conf/cvpr/ZhouKLOT16 fatcat:4mmwelc4xbgr5gf4erobt5cmpi

Learning Deep Features for Discriminative Localization [article]

Bolei Zhou and Aditya Khosla and Agata Lapedriza and Aude Oliva and Antonio Torralba
2015 arXiv   pre-print
apparent simplicity of global average pooling, we are able to achieve 37.1% top-5 error for object localization on ILSVRC 2014, which is remarkably close to the 34.2% top-5 error achieved by a fully supervised  ...  Weakly supervised text detector: We train a weakly supervised text detector using 350 Google StreetView images containing text from the SVT dataset [25] as the positive set and randomly sampled images  ...  Figure 11 . 11 Learning a weakly supervised text detector. The text is accurately detected on the image even though our network is not trained with text or any bounding box annotations.  ... 
arXiv:1512.04150v1 fatcat:zkorupnj6jhi7n2c7lswueg5gq

Movable-Object-Aware Visual SLAM via Weakly Supervised Semantic Segmentation [article]

Ting Sun, Yuxiang Sun, Ming Liu, Dit-Yan Yeung
2019 arXiv   pre-print
In this paper we propose to adopt a novel weakly-supervised semantic segmentation method. The segmentation mask is obtained from a CNN pre-trained with image-level class labels only.  ...  Moving objects can greatly jeopardize the performance of a visual simultaneous localization and mapping (vSLAM) system which relies on the static-world assumption.  ...  To the best of our knowledge, this is the first work that adopts weakly supervised semantic segmentation CNN for dynamic objects aware vSLAM. [27] .  ... 
arXiv:1906.03629v2 fatcat:ebynlvgfjfdtth7hnao3guq73a

FeTaQA: Free-form Table Question Answering

Linyong Nan, Chiachun Hsieh, Ziming Mao, Xi Victoria Lin, Neha Verma, Rui Zhang, Wojciech Kryściński, Nick Schoelkopf, Riley Kong, Xiangru Tang, Mutethia Mutuma, Ben Rosand (+5 others)
2022 Transactions of the Association for Computational Linguistics  
reasoning and integration of information pieces retrieved from a structured knowledge source.  ...  However, restricted by their short-form answers, these datasets fail to include question–answer interactions that represent more advanced and naturally occurring information needs: questions that ask for  ...  In the real world, knowledge sources take a variety of forms, including unstructured texts (documents, passages, or conversations), structured knowledge bases, and semi-structured tables, each requiring  ... 
doi:10.1162/tacl_a_00446 fatcat:iuumocwzzzel5ltsznv5y4u6yq

Learning by Fixing: Solving Math Word Problems with Weak Supervision [article]

Yining Hong, Qing Li, Daniel Ciao, Siyuan Huang, Song-Chun Zhu
2021 arXiv   pre-print
In this paper, we address this issue by introducing a weakly-supervised paradigm for learning MWPs.  ...  Experimental results on the Math23K dataset show the proposed LBF framework significantly outperforms reinforcement learning baselines in weakly-supervised learning.  ...  Ethical Impact The presented work should be categorized as research in the field of weakly-supervised learning and abductive reasoning.  ... 
arXiv:2012.10582v2 fatcat:p2g7gnfhurhf7imapt6wm3euhi
« Previous Showing results 1 — 15 out of 4,750 results