42,649 Hits in 2.5 sec

Unsupervised Batch Normalization

Mustafa Taha Kocyigit, Laura Sevilla-Lara, Timothy M. Hospedales, Hakan Bilen
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)  
We propose using such unlabeled examples to calculate batch normalization statistics, which we call Unsupervised Batch Normalization (UBN).  ...  Batch Normalization is a widely used tool in neural networks to improve the generalization and convergence of training.  ...  + (1 − λ)σ(x c ) (3) Unsupervised Batch Normalization UBN is based on updating first the batch statistics than the weights.  ... 
doi:10.1109/cvprw50498.2020.00467 dblp:conf/cvpr/KocyigitSHB20 fatcat:32jtqxhmn5d4ze2y5cxxtbwefu

Unsupervised Out-of-Distribution Detection with Batch Normalization [article]

Jiaming Song and Yang Song and Stefano Ermon
2019 arXiv   pre-print
To address these issues, we propose a new method that leverages batch normalization.  ...  We argue that batch normalization for generative models challenges the traditional i.i.d. data assumption and changes the corresponding maximum likelihood objective.  ...  For a batch of inputs (Ioffe & Szegedy, 2015) performs normalization over the inputs followed by a parametrized affine transformation: z = {z i } b i=1 of batch size b, batch normalization BatchNorm(  ... 
arXiv:1910.09115v1 fatcat:dggjj4ldwng2xfwqqqkkqppene

Unsupervised Model Drift Estimation with Batch Normalization Statistics for Dataset Shift Detection and Model Selection [article]

Wonju Lee, Seok-Yong Byun, Jooeun Kim, Minje Park, Kirill Chechil
2021 arXiv   pre-print
In this paper, we propose a novel method of model drift estimation by exploiting statistics of batch normalization layer on unlabeled test data.  ...  However, it is less possible to annotate or inspect newly streamed data by humans, and thus it is desired to measure model drift at inference time in an unsupervised manner.  ...  In this paper, we presented a novel method of estimating model drift by exploiting the statistics of batch normalization layer in an unsupervised manner.  ... 
arXiv:2107.00191v1 fatcat:hjy7ohbqtfbchorjlcnjyclxua

A Domain Agnostic Normalization Layer for Unsupervised Adversarial Domain Adaptation [article]

Rob Romijnders, Panagiotis Meletis, Gijs Dubbelman
2018 arXiv   pre-print
We propose a normalization layer for unsupervised domain adaption in semantic scene segmentation.  ...  Therefore, we propose a novel Domain Agnostic Normalization layer and thereby unlock the benefits of normalization layers for unsupervised adversarial domain adaptation.  ...  The retrieval curve in Figure 4 shows another view on the effects of batch normalization in unsupervised domain adaptation.  ... 
arXiv:1809.05298v1 fatcat:3ymqfuvb6vaqtfpgjiuayycdzi

Modified Autoencoder Training and Scoring for Robust Unsupervised Anomaly Detection in Deep Learning

Nicholas Merrill, Azim Eskandarian
2020 IEEE Access  
AEs are trained on the assumption that abnormal inputs will produce higher reconstruction errors than normal ones.  ...  In practice, however, this assumption is unreliable in the unsupervised case, where the training data may contain anomalous examples.  ...  Furthermore, difficult normal examples that fall above the threshold in one-batch, because of the stochastic draws, may appear below the threshold in another batch.  ... 
doi:10.1109/access.2020.2997327 fatcat:dykw6htq6ncupc7jclldlmdzwu

Incremental Real-Time Personalization in Human Activity Recognition Using Domain Adaptive Batch Normalization [article]

Alan Mazankiewicz, Klemens Böhm, Mario Bergés
2020 arXiv   pre-print
Previous work has addressed this challenge by personalizing general recognition models to the unique motion pattern of a new user in a static batch setting.  ...  Our work addresses this challenges by proposing an unsupervised online domain adaptation algorithm. Both classification and personalization happen continuously and incrementally in real-time.  ...  Domain Adaptive Batch Normalization Batch normalization can be applied to unsupervised single-source-single-target domain adaptation by a simple change in the testing phase, as proposed by [8] .  ... 
arXiv:2005.12178v1 fatcat:6twxy2u4grbnnldc4l2rnsysni

Oncocytoma-Related Gene Signature to Differentiate Chromophobe Renal Cancer and Oncocytoma Using Machine Learning

Khaled Bin Satter, Paul Minh Huy Tran, Lynn Kim Hoang Tran, Zach Ramsey, Katheine Pinkerton, Shan Bai, Natasha M. Savage, Sravan Kavuri, Martha K. Terris, Jin-Xiong She, Sharad Purohit
2022 Cells  
The transcriptomic differences were identified with unsupervised learning in the discovery dataset (97.8% accuracy) with density based UMAP (DBU).  ...  (RO) and normal kidney tissue arrays (N) before batch effect correction.  ...  (chRCC) and renal oncocytoma (RO) and normal kidney tissue arrays (N) before batch effect correction.  ... 
doi:10.3390/cells11020287 pmid:35053403 pmcid:PMC8774230 fatcat:lzubkmknlvclnc5ytnjmrh5kcu

Just DIAL: DomaIn Alignment Layers for Unsupervised Domain Adaptation [article]

Fabio Maria Carlucci and Lorenzo Porzi and Barbara Caputo and Elisa Ricci and Samuel Rota Bulò
2017 arXiv   pre-print
: Batch normalization.  ...  Laplace Batch normalization.  ... 
arXiv:1702.06332v3 fatcat:ap5s7kza4zgbnbvntqfoob6pum

Cold Start Approach for Data-Driven Fault Detection

Mihajlo Grbovic, Weichang Li, Niranjan A Subrahmanya, Usadi, Slobodan Vucetic
2013 IEEE Transactions on Industrial Informatics  
The proposed fusion model performed better on both unseen and seen faults than the stand-alone unsupervised and supervised models.  ...  In this paper we study this often overlooked cold start learning problem in data-driven fault detection, where in the beginning only normal operation data are available and faulty operation data become  ...  In the second scenario, the batches are partially labeled. PCA was used as the unsupervised model.  ... 
doi:10.1109/tii.2012.2231870 fatcat:spawvsr4y5cpxasj7mcsuedvaq

Generative Cooperative Learning for Unsupervised Video Anomaly Detection [article]

Muhammad Zaigham Zaheer, Arif Mahmood, Muhammad Haris Khan, Mattia Segu, Fisher Yu, Seung-Ik Lee
2022 arXiv   pre-print
In essence, both networks get trained in a cooperative fashion, thereby allowing unsupervised learning.  ...  Consistent improvement over the existing state-of-the-art unsupervised and OCC methods corroborate the effectiveness of our approach.  ...  Note that, the term 'unsupervised' in literature often refers to OCC approaches which assume all normal training data [10, 36, 62] .  ... 
arXiv:2203.03962v1 fatcat:sq2nycz2ubhzlpn7biurdhb7xi

Clustering and Unsupervised Anomaly Detection with L2 Normalized Deep Auto-Encoder Representations [article]

Caglar Aytekin, Xingyang Ni, Francesco Cricri, Emre Aksu
2018 arXiv   pre-print
We also propose a clustering based unsupervised anomaly detection method using l2 normalized deep auto-encoder representations. We show the effect of l2 normalization on anomaly detection accuracy.  ...  With the advent of deep learning, there is an increasing interest in learning deep unsupervised representations for clustering analysis.  ...  We have shown that the proposed l 2 normalization constraint drastically increases the anomaly detection method's performance.  ... 
arXiv:1802.00187v1 fatcat:hlvka7di75dabaheyfxpqg5jty

Supervised normalization of microarrays

Brigham H. Mecham, Peter S. Nelson, John D. Storey
2010 Computer applications in the biosciences : CABIOS  
as batch or array processing date.  ...  It is intuitively clear that true biological signal and confounding factors need to be simultaneously considered when performing normalization.  ...  Also, suppose that the arrays were processed in two separate batches. Unsupervised methods ignore the treatment and batch variables when performing the normalization.  ... 
doi:10.1093/bioinformatics/btq118 pmid:20363728 pmcid:PMC2865860 fatcat:rimcgcqfdnh3jcs5cidsd4uz5u

MixNorm: Test-Time Adaptation Through Online Normalization Estimation [article]

Xuefeng Hu, Gokhan Uzunbas, Sirius Chen, Rui Wang, Ashish Shah, Ram Nevatia, Ser-Nam Lim
2021 arXiv   pre-print
proposed method significantly outperforms the State-Of-The-Art in the newly proposed settings in Test-Time Adaptation Task, and also demonstrates improvements in various other settings such as Source-Free Unsupervised  ...  Unlike the previous methods that require a large batch of single distribution during test time to calculate stable batch-norm statistics, our method avoid any dependency on large online batches and is  ...  TTT performs a single step gradient update over the batch-norm parameters after each prediction, with an unsupervised loss defined by rotation prediction task.  ... 
arXiv:2110.11478v1 fatcat:po2tg35wpfciroi434rv6gkrx4

Shuffle and Learn: Unsupervised Learning using Temporal Order Verification [article]

Ishan Misra and C. Lawrence Zitnick and Martial Hebert
2016 arXiv   pre-print
We formulate our method as an unsupervised sequential verification task, i.e., we determine whether a sequence of frames from a video is in the correct temporal order.  ...  Control for batch normalization: We use batch normalization [59] for training our triple-Siamese network.  ...  Since the other baseline methods from [60] predate the batch normalization method, we first re-trained the baseline models using batch normalization.  ... 
arXiv:1603.08561v2 fatcat:ydbpme3cdfhlvalrbrkzs4gjyu

Bearing Fault Diagnosis Based on the Switchable Normalization SSGAN with 1-D Representation of Vibration Signals as Input

Dongdong Zhao, Feng Liu, He Meng
2019 Sensors  
Switchable normalization and semi-supervised learning are introduced to solve the above obstacles in this paper, which proposes a novel bearing fault diagnosis method based on switchable normalization  ...  Batch normalization (BN): normalizes the NHW on the batch; Instance normalization (IN): normalizes the HW on the image pixels; Layer normalization (LN): normalizes the CHW in the channel direction; Group  ...  Batch normalization (BN): normalizes the NHW on the batch; Instance normalization (IN): normalizes the HW on the image pixels; Layer normalization (LN): normalizes the CHW in the channel direction; Group  ... 
doi:10.3390/s19092000 fatcat:eb333gm5qbe5dboygy2ikm7oca
« Previous Showing results 1 — 15 out of 42,649 results