1,128 Hits in 7.9 sec

Intimate Learning: A Novel Approach for Combining Labelled and Unlabelled Data

Zhongmin Shi, Anoop Sarkar
2005 International Joint Conference on Artificial Intelligence  
The method is tested on a Web information extraction task of learning course names from web pages in which we use very few labelled items as seed data (10 web pages) and combine with an unlabelled set  ...  The overall performance improved the precision/recall from 3.11%/0.31% for a baseline EM-based method to 44.7%/44.1% for intimate learning.  ...  labelled and unlabelled data.  ... 
dblp:conf/ijcai/ShiS05 fatcat:dydi6hevv5dunmjqrmhmfnnoay

Active Learning for Anomaly and Rare-Category Detection

Dan Pelleg, Andrew W. Moore
2004 Neural Information Processing Systems  
We introduce a novel active-learning scenario in which a user wants to work with a learning algorithm to identify useful anomalies.  ...  The challenge is thus to identify "rare category" records in an unlabeled noisy set with help (in the form of class labels) from a human expert who has a small budget of datapoints that they are prepared  ...  Our application presents multiple indicators to help a user spot anomalous data, as well as controls for labeling points and adding classes. The application will be described in a companion paper.  ... 
dblp:conf/nips/PellegM04 fatcat:72rjrc4mtbdajgitslsvsdnb7m

Score Function Features for Discriminative Learning: Matrix and Tensor Framework [article]

Majid Janzamin, Hanie Sedghi, Anima Anandkumar
2014 arXiv   pre-print
In this paper, we consider a novel class of matrix and tensor-valued features, which can be pre-trained using unlabeled samples.  ...  Thus, we present a novel framework for employing generative models of the input for discriminative learning.  ...  Janzamin thanks Rina Panigrahy for useful discussions. M. Janzamin is supported by NSF Award CCF-1219234. H. Sedghi is supported by ONR Award N00014-14-1-0665. A.  ... 
arXiv:1412.2863v2 fatcat:44tmovmirbcmrfm3cm2ohipocq

Learning to Abstain from Binary Prediction [article]

Akshay Balsubramani
2016 arXiv   pre-print
We give an algorithm for learning a classifier in this setting which trades off its errors with abstentions in a minimax optimal manner, is as efficient as linear learning and prediction, and is demonstrably  ...  A binary classifier capable of abstaining from making a label prediction has two goals in tension: minimizing errors, and avoiding abstaining unnecessarily often.  ...  A primary contribution of this work is to describe the Pareto frontier completely for a very general semisupervised learning scenario, in which i.i.d. data are available in both labeled and unlabeled form  ... 
arXiv:1602.08151v2 fatcat:23ikvhvqhvfxlaf23uqn6ppwo4

Simple strategies for semi-supervised feature selection

Konstantinos Sechidis, Gavin Brown
2017 Machine Learning  
If we have some binary labelled data and some unlabelled, we could assume the unlabelled data are all positives, or assume them all negatives.  ...  However, with theoretical and empirical studies, we show they provide powerful results for feature selection, via hypothesis testing and feature ranking.  ...  Figure 11 shows that the approaches that use both labelled and unlabelled data-MINT, Figure 12 shows that the approaches that use both labelled and unlabelled data-MINT, Semi-MIM, Semi-JMI-outperform  ... 
doi:10.1007/s10994-017-5648-2 pmid:31983804 pmcid:PMC6954040 fatcat:ze4ljb4xlng7zdw5btvdcz52ue

Zero-Shot Learning with Knowledge Enhanced Visual Semantic Embeddings [article]

Karan Sikka, Jihua Huang, Andrew Silberfarb, Prateeth Nayak, Luke Rohrer, Pritish Sahu, John Byrnes, Ajay Divakaran, Richard Rohwer
2020 arXiv   pre-print
We introduce two key novelties for improved learning: (1) enforcement of rules for a group instead of a single concept to take into account class-wise relationships, and (2) confidence margins inside logical  ...  operators that enable implicit curriculum learning and prevent premature overfitting.  ...  Acknowledgement The authors would like to acknowledge Karen Myers, Bill Mark, and Rodrigo Braz for helpful discussions.  ... 
arXiv:2011.10889v1 fatcat:irbppidpyrggzmqrficji23ida

Functional Nanomaterials Design in the Workflow of Building Machine-Learning Models [article]

Zhexu Xi
2021 arXiv   pre-print
With enormous potentiality to tackle more real-world problems, ML provides a more comprehensive insight into combinations with molecules/materials under the fundamental procedures for constructing ML models  ...  and application of novel functional materials, especially at the nanometre scale.  ...  In semi-supervised learning, only limited sets of feature (labeled) data can be output from excessive amounts of unlabeled data.  ... 
arXiv:2108.13171v1 fatcat:6wdmhygabrccvnmlnapqz4ajce

A Comparative Study of Methods for Transductive Transfer Learning

Andrew Arnold, Ramesh Nallapati, William W. Cohen
2007 Seventh IEEE International Conference on Data Mining Workshops (ICDMW 2007)  
Previous work has studied the supervised version of this problem in which labeled data from both source and target domains are available for training.  ...  In this work, however, we study the more challenging problem of unsupervised transductive transfer learning, where no labeled data from the target domain are available at training time, but instead, unlabeled  ...  Specifically, a hyperplane is trained on the labeled source data and then used to classify the unlabeled testing data.  ... 
doi:10.1109/icdmw.2007.109 dblp:conf/icdm/ArnoldNC07 fatcat:nef4zkreorgk3kqimgo7mbiezu

Deep Learning for Sensor-based Human Activity Recognition: Overview, Challenges and Opportunities [article]

Kaixuan Chen, Dalin Zhang, Lina Yao, Bin Guo, Zhiwen Yu, Yunhao Liu
2021 arXiv   pre-print
In this study, we present a survey of the state-of-the-art deep learning methods for sensor-based human activity recognition.  ...  We first introduce the multi-modality of the sensory data and provide information for public datasets that can be used for evaluation in different challenge tasks.  ...  The classifier learns from a small amount of labeled data, selects one or a set of the most useful unlabeled samples via query strategy, ask the annotator for true labels, and utilize the new labels for  ... 
arXiv:2001.07416v2 fatcat:km2b3xn4sngtxgkdck6ymlmu3m

Feedback Coding for Active Learning [article]

Gregory Canal, Matthieu Bloch, Christopher Rozell
2021 arXiv   pre-print
The iterative selection of examples for labeling in active machine learning is conceptually similar to feedback channel coding in information theory: in both tasks, the objective is to seek a minimal sequence  ...  We evaluate APM on a variety of datasets and demonstrate learning performance comparable to existing active learning methods, at a reduced computational cost.  ...  Acknowledgements We thank the reviewers for their useful feedback and comments, as well as Rob Nowak, John Lee, and other colleagues for insightful discussions.  ... 
arXiv:2103.00654v1 fatcat:iwu3rivovnaxfahpzxwbfhmzua

Machine learning based hyperspectral image analysis: A survey [article]

Utsav B. Gewali, Sildomar T. Monteiro, Eli Saber
2019 arXiv   pre-print
Therefore, a solid understanding of machine learning techniques have become essential for remote sensing researchers and practitioners.  ...  Machine learning algorithms due to their outstanding predictive power have become a key tool for modern hyperspectral image analysis.  ...  Active learning [272] iteratively selects examples from the unlabeled data for manual labeling, and adds them to the labeled training set.  ... 
arXiv:1802.08701v2 fatcat:bfi6qkpx2bf6bowhyloj2duugu

Transductive hyperspectral image classification: toward integrating spectral and relational features via an iterative ensemble system

Annalisa Appice, Pietro Guccione, Donato Malerba
2016 Machine Learning  
The transductive inference paradigm allows us to reduce the inference error for the given set of unlabeled data, as sparsely labeled pixels are learned by accounting for both labeled and unlabeled information  ...  We propose a novel algorithm to assign a class to each pixel of a sparsely labeled hyperspectral image. It integrates the spectral information and the spatial correlation through an ensemble system.  ...  Acknowledgments We would like to acknowledge the support of the European Commission through the project MAESTRA-Learning from Massive, Incompletely annotated, and Structured Data (Grant Number ICT-2013  ... 
doi:10.1007/s10994-016-5559-7 fatcat:ims577xfwrgspcyntnngwztfxa

Relieving the Plateau: Active Semi-Supervised Learning for a Better Landscape [article]

Seo Taek Kong, Soomin Jeon, Jaewon Lee, Hongseok Lee, Kyu-Hwan Jung
2021 arXiv   pre-print
Given the accessible pool of unlabeled data in pool-based AL, it seems natural to use SSL when training and AL to update the labeled set; however, algorithms designed for their combination remain limited  ...  Active learning (AL) selects unlabeled instances to be annotated by a human-in-the-loop in hopes of better performance with less labeled data.  ...  First, CRC scores the value of labeling data over a combination of already-labeled data and candidate unlabeled data, explicitly consolidating existing labeled data.  ... 
arXiv:2104.03525v1 fatcat:cdxrqus7bbeofaxa67ocrcqz6i

Deep Learning Based Pain Treatment

Tarun Jaiswal, Sushma Jaiswal
2019 International Journal of Trend in Scientific Research and Development  
Among machine learning methods, a subset has so far been applied to pain research-related problems, SVMs, regression models, deep learning and several kinds of neural networks so far most often revealed  ...  Indeed, the application of machine learning for pain investigationassociated non-imaging problems has been mentioned in publications in scientific journals since 1940-2018.  ...  This paper, introduce the novel concept of Graph Convolutional Networks (GCN) for brain analysis in populations, combining imaging and non-imaging data.  ... 
doi:10.31142/ijtsrd23639 fatcat:tqg4u3tkgjhmjpya67g3lnewwu

Deep Learning applications for COVID-19

Connor Shorten, Taghi M. Khoshgoftaar, Borko Furht
2021 Journal of Big Data  
These limitations include Interpretability, Generalization Metrics, Learning from Limited Labeled Data, and Data Privacy.  ...  We describe how each of these applications vary with the availability of big data and how learning tasks are constructed.  ...  Acknowledgements We would like to thank the reviewers in the Data Mining and Machine Learning Laboratory at Florida Atlantic University.  ... 
doi:10.1186/s40537-020-00392-9 pmid:33457181 pmcid:PMC7797891 fatcat:aokxo63z2rhdpfxo3egyf3xpcm
« Previous Showing results 1 — 15 out of 1,128 results