Filters








2,450 Hits in 12.3 sec

Embracing noise to improve cross-batch prediction accuracy

Chuan Hock Koh, Limsoon Wong
2012 BMC Systems Biology  
However, the results were not very encouraging, as prediction performance did not always improve. In fact, in up to 20% of the cases, prediction accuracy was reduced.  ...  Recently, a prominent study was published on how batch effects removal techniques could potentially improve microarray prediction performance.  ...  Thus, instead of attempting to estimate noise due to batch effects, we embrace it by using rank values rather than absolute values.  ... 
doi:10.1186/1752-0509-6-s2-s3 pmid:23282067 pmcid:PMC3521182 fatcat:ahefyzur6jfcxov7ful5rszkni

Deep Hierarchical Product Classification Based on Pre-Trained Multilingual Knowledge

Wen Zhang, Yanbin Lu, Bella Dubrov, Zhi Xu, Shang Shang, Emilio Maldonado
2021 IEEE Data Engineering Bulletin  
Improvement(+)/Downgrade(-) of the department/leaf level accuracy is reported.  ...  Machine learning based systems often suffer from poor data quality, such as incomplete item descriptions, adversarial noise in the training data, etc., causing low precision/recall of predictions.  ...  efficiently predict product categories in both department and leaf category levels; 3) We propose several training strategies to further improve classification accuracy.  ... 
dblp:journals/debu/ZhangLDXSM21 fatcat:24kmexajdzh5jdro6njfziipji

Implicit Generative Modeling of Random Noise during Training for Adversarial Robustness [article]

Priyadarshini Panda, Kaushik Roy
2019 arXiv   pre-print
We introduce a Noise-based prior Learning (NoL) approach for training neural networks that are intrinsically robust to adversarial attacks.  ...  We find that the implicit generative modeling of random noise with the same loss function used during posterior maximization, improves a model's understanding of the data manifold furthering adversarial  ...  prediction.  ... 
arXiv:1807.02188v4 fatcat:z2idp5pbirf3dgwt52emsnye3m

A Generalized Supervised Contrastive Learning Framework [article]

Jaewon Kim, Jooyoung Chang, Sang Min Park
2022 arXiv   pre-print
to the supervised context and outperformed cross-entropy on various datasets on ResNet.  ...  ResNet-50 trained in GenSCL with Mixup-Cutmix and KD achieves state-of-the-art accuracies of 97.6% and 84.7% on CIFAR10 and CIFAR100 without external data, which significantly improves the results reported  ...  In addition to cross-entropy, Kullback-Leibler divergence between student's prediction and teacher's prediction can boost the training of the student.  ... 
arXiv:2206.00384v1 fatcat:57tuslgopvcfjaapwevfx7ndju

Joint Attention-Driven Domain Fusion and Noise-Tolerant Learning for Multi-Source Domain Adaptation [article]

Tong Xu, Wu Ning, Chunyan Lyu, Kejun Wang
2022 arXiv   pre-print
In light of this, we propose an approach that integrates Attention-driven Domain fusion and Noise-Tolerant learning (ADNT) to address the two issues mentioned above.  ...  Through this approach, the discriminability of the features can also be significantly improved while the domain discrepancy is reduced.  ...  And based on [15] , [26] introduced a label shifting strategy as a way to improve the accuracy of low confidence predictions. [14] applied F-norm and L 21 -norm regularization operations to firmly  ... 
arXiv:2208.02947v1 fatcat:2bm7ml4zwnedjejz3abihf3txy

Embracing Ambiguity: Shifting the Training Target of NLI Models [article]

Johannes Mario Meissner, Napat Thumwanit, Saku Sugawara, Akiko Aizawa
2021 arXiv   pre-print
While many research works do not pay much attention to this fact, several recent efforts have been made to acknowledge and embrace the existence of ambiguity, such as UNLI and ChaosNLI.  ...  how to capture linguistic ambiguity.  ...  Previously, it was common to disregard it as noise or as a sign of poor quality data.  ... 
arXiv:2106.03020v1 fatcat:zvs5ple4rzdqhmccizecgba6pa

Data Fusion for Enhanced Fermentation Process Tracking

Shengnan Yu, Gary Montague, Elaine Martin
2010 IFAC Proceedings Volumes  
The methodologies are applied to data from an industrial fermentation process and it is shown that the data fusion method results in a 50% improvement in the Root Mean Square Error of Cross Validation  ...  Aiming to overcome the limitations of sequential modeling and to compare model accuracy, a novel data fusion methodology based on Partial Least Squares, weighted multivariate calibration, is introduced  ...  The TSB is an executive body established by the Government to drive innovation.  ... 
doi:10.3182/20100705-3-be-2011.00007 fatcat:563nz7m4n5c2fmik7g757eadzu

Exploring and Improving Robustness of Multi Task Deep Neural Networks via Domain Agnostic Defenses [article]

Kashyap Coimbatore Murali
2020 arXiv   pre-print
Liu et al., have shown that the Multi-Task Deep Neural Network, due to the regularization effect produced when training as a result of its cross task data, is more robust than a vanilla BERT model trained  ...  Finally, we propose a domain agnostic defense which restores the model's accuracy (36.75% and 25.94% respectively) as opposed to a general-purpose defense or an off-the-shelf spell checker.  ...  ) [20] and commercial spell checkers for all the noises that it was tested on (adding -3.52% absolute improvement, deleting -13.89% absolute improvement, and jumbling -41.85% absolute improvement).  ... 
arXiv:2001.05286v1 fatcat:4ttls5pkcraatclu3ycofdr7ka

Noise-Tolerant Learning for Audio-Visual Action Recognition [article]

Haochen Han, Qinghua Zheng, Minnan Luo, Kaiyao Miao, Feng Tian, Yan Chen
2022 arXiv   pre-print
To reduce the influence of noisy correspondence, we propose a cross-modal noise estimation component to adjust the consistency between different modalities.  ...  Recently, video recognition is emerging with the help of multi-modal learning, which focuses on integrating multiple modalities to improve the performance or robustness of a model.  ...  In order to avoid trivial solution that all samples are predicted to same label, one constraint is added to partition the batch in equal size: B i=1 q(y|a i ) = B K and B i=1 q(y|v i ) = B K . ( 10 ) The  ... 
arXiv:2205.07611v2 fatcat:lzf66dhinrfxjgo3lrw4c7lb2a

WiPg: Contactless Action Recognition Using Ambient Wi-Fi Signals

Zhanjun Hao, Juan Niu, Xiaochao Dang, Zhiqiang Qiao
2022 Sensors  
the wireless signal environment of human activity usually carries a large amount of human-related information, the motion-recognition model trained for a specific person usually does not work well in predicting  ...  The experimental results show that the average correct rate of WiPg can reach 92.7% for recognition of the 14 yoga poses, and WiPg realizes "cross-personnel" movement recognition with excellent recognition  ...  Cross-domain gesture recognition is realized by the generalization technique, and the accuracy of gesture recognition based on CSI was improved to 92.4% in the cross-environment situation.  ... 
doi:10.3390/s22010402 pmid:35009943 pmcid:PMC8749714 fatcat:nsgp7z73fbb4zpt2hzp25hxlmm

Nodule Detection with Convolutional Neural Network Using Apache Spark and GPU Frameworks

Nikitha Johnsirani Venkatesan, Dong Ryeol Shin, Choon Sung Nam
2021 Applied Sciences  
High radiation dose provokes cancer, which forces experts to use limited radiation. Using abrupt radiation generates noise in CT scans.  ...  We propose an optimal Convolutional Neural Network model in which Gaussian noise is removed for better classification and increased training accuracy.  ...  The batch size of the data and the learning rate should be fixed according to the training accuracy to increase the throughput.  ... 
doi:10.3390/app11062838 fatcat:axcyrh5aunga7i5iz465dpd7fe

Cataract Eye Prediction using Machine Learning

Shruthi Bhat, Som Mosalagi, Tejal Bhalerao, Pushpak Katkar, Rahul Pitale
2020 International Journal of Computer Applications  
This would rather ensure that people belonging to remote areas need not reach out to ophthalmologists, just to check whether the person is facing a cataract problem or not.  ...  The challenge is to detect cataract using the normal lens images at an early stage thus allowing people to test for cataract themselves.  ...  The second one is used to increase the number of images available to improve accuracy. [6] Cataract must be detected as soon as possible so that it will be easy to prevent it by getting more and turning  ... 
doi:10.5120/ijca2020920441 fatcat:lim4tfpeefetfoicdvvijekdri

Deep learning regression for inverse quantum scattering [article]

A. C. Maioli
2020 arXiv   pre-print
A investigation with noisy data is presented and it is observed that the neural network is useful to predict the potential parameters.  ...  A step-by-step method is provided in order to obtain the potential parameters. A circular boundary-wall potential was chosen to exemplify the method.  ...  This separation is important to check the accuracy of the network.  ... 
arXiv:2009.09944v1 fatcat:2q744b5b5zaozdbeze4lsztfdq

In Defense of the Triplet Loss Again: Learning Robust Person Re-Identification with Fast Approximated Triplet Loss and Label Distillation [article]

Ye Yuan, Wuyang Chen, Yang Yang, Zhangyang Wang
2019 arXiv   pre-print
accuracy, efficiency, robustness, and direct transferability to unseen datasets.  ...  Moreover, the abundance of label noise and outliers in ReID datasets may also put the margin-based loss in jeopardy.  ...  with a loss that is less sensitive to label noise, e.g., cross entropy.  ... 
arXiv:1912.07863v2 fatcat:z3z3wipixnhxvncthhn3s4ytxq

Material Classification with a Transfer Learning based Deep Model on an imbalanced Dataset using an epochal Deming-Cycle-Methodology

Marco Klaiber
2022 ELCVIA Electronic Letters on Computer Vision and Image Analysis  
In addition, care was taken to achieve a balance between accuracy and robustness with respect to the model.  ...  The deep learning model evaluated achieved 91.54% accuracy with the dataset used and set new standards with the method applied.  ...  for improving predictive performance in these cases [40, 41] .  ... 
doi:10.5565/rev/elcvia.1517 doaj:1729c7e603284b428debb7adf4a063b5 fatcat:tcxd3f5bhzdqdidau26mdpcsli
« Previous Showing results 1 — 15 out of 2,450 results