Filters








7,662 Hits in 1.8 sec

Towards Propagation Uncertainty: Edge-enhanced Bayesian Graph Convolutional Networks for Rumor Detection [article]

Lingwei Wei, Dou Hu, Wei Zhou, Zhaojuan Yue, Songlin Hu
2021 arXiv   pre-print
Detecting rumors on social media is a very critical task with significant implications to the economy, public health, etc.  ...  Towards this issue, this paper makes the first attempt to explore propagation uncertainty for rumor detection.  ...  to optimize the model with unlabeled latent relations. • Experiments on three real-world benchmark datasets demonstrate the effectiveness of our model on both rumor detection and early rumor detection  ... 
arXiv:2107.11934v1 fatcat:bosu37bur5dtrctae5y4sa233m

Entropy-Based Latent Structured Output Prediction

Diane Bouchacourt, Sebastian Nowozin, M. Pawan Kumar
2015 2015 IEEE International Conference on Computer Vision (ICCV)  
; and (ii) those that predict the output variables by minimizing an entropy-based uncertainty measure over the latent space.  ...  Recently several generalizations of the popular latent structural SVM framework have been proposed in the literature.  ...  latent space with the use of an entropy-based uncertainty measure.  ... 
doi:10.1109/iccv.2015.334 dblp:conf/iccv/BouchacourtNK15 fatcat:r3rvkdnn5zhillkorrqk7qotli

Dense Uncertainty Estimation [article]

Jing Zhang, Yuchao Dai, Mochu Xiang, Deng-Ping Fan, Peyman Moghadam, Mingyi He, Christian Walder, Kaihao Zhang, Mehrtash Harandi, Nick Barnes
2021 arXiv   pre-print
Bayesian Neural Networks) or including latent variables (i.e. generative models) to explore the contribution of latent variables for model predictions, leading to stochastic predictions during testing.  ...  In this paper, we investigate stochastic neural networks and uncertainty estimation techniques to achieve both accurate deterministic prediction and reliable uncertainty estimation.  ...  For both salient object detection and camouflaged object detection, we adopt the structure-aware loss function [68] .  ... 
arXiv:2110.06427v1 fatcat:a4f3tcjyz5ftdokazwntegqojm

A General Divergence Modeling Strategy for Salient Object Detection [article]

Xinyu Tian, Jing Zhang, Yuchao Dai
2021 arXiv   pre-print
Salient object detection is subjective in nature, which implies that multiple estimations should be related to the same input image.  ...  Although latent variable model based stochastic prediction network exists to model the prediction variants, the latent space based on the single clean saliency annotation is less reliable in exploring  ...  Task-related loss function: The widely used loss functions for saliency detection include: 1) binary cross-entropy loss, 2) boundary IOU loss [30] , and 3) structure-aware loss [54] , which is a weighted  ... 
arXiv:2111.11827v1 fatcat:7yhcithrxfabzofvcqkyhi36pa

Uncertainty Inspired RGB-D Saliency Detection [article]

Jing Zhang, Deng-Ping Fan, Yuchao Dai, Saeed Anwar, Fatemeh Saleh, Sadegh Aliakbarian, Nick Barnes
2020 arXiv   pre-print
We propose the first stochastic framework to employ uncertainty for RGB-D saliency detection by learning from the data labeling process.  ...  To infer the latent variable, we introduce two different solutions: i) a Conditional Variational Auto-encoder with an extra encoder to approximate the posterior distribution of the latent variable; and  ...  To prove that our model can also works well with basic cross-entropy loss, we designed another model with cross-entropy loss used instead of the structure-aware loss, and show performance as "CE".  ... 
arXiv:2009.03075v1 fatcat:acdr5aepnjdynhgx35i2b4brj4

Probabilistic Parameter Selection for Learning Scene Structure from Video

M.D. Breitenstein, E. Sommerlade, B. Leibe, L. Van Gool, I. Reid
2008 Procedings of the British Machine Vision Conference 2008  
Our approach is based on an entropy modelling framework, which allows to simultaneously adapt the detector parameters, such that the expected information gain about the scene structure is maximised.  ...  To resolve this tradeoff, we propose a method for informed parameter selection which minimises the expected uncertainty of the scene structure estimate based on an entropy framework.  ...  Its purpose is to integrate information from object detector responses in order to simultaneously estimate local scene structure and to provide the entropy framework with a means of predicting what effects  ... 
doi:10.5244/c.22.32 dblp:conf/bmvc/BreitensteinSLGR08 fatcat:qyf5jorkl5dmtfoykywtklqtxi

Asking without Telling: Exploring Latent Ontologies in Contextual Representations [article]

Julian Michael, Jan A. Botha, Ian Tenney
2020 arXiv   pre-print
To investigate this, we introduce latent subclass learning (LSL): a modification to existing classifier-based probing methods that induces a latent categorization (or ontology) of the probe's inputs.  ...  If so, how is this structure encoded?  ...  diversity and uncertainty, while the instance-level entropy loss drives them down.  ... 
arXiv:2004.14513v2 fatcat:4nj2vqn4rnepfdvjnqgny7umai

Computation in Complex Networks

Clara Pizzuti, Annalisa Socievole
2021 Entropy  
Acknowledgments: We express our thanks to the authors of the above contributions, and to the journal Entropy and MDPI for their support during this work.  ...  The first experiment assessed the impact of the Bayesian network structure on the entropy of the model.  ...  The second compared the entropy of the posterior distribution of the class variable obtained from the different structures.  ... 
doi:10.3390/e23020192 pmid:33562478 fatcat:g2ad4eiupnaydp653k7hvzeqyu

Uncertainty Evaluation in Multistage Assembly Process Based on Enhanced OOPN

Yubing Huang, Wei Dai, Weiping Mou, Yu Zhao
2018 Entropy  
Finally, this work analyzed the assembly process on the basis of the uncertainty of the assembly structure and the variables of the assembly process.  ...  The definition of entropy in physics was applied to characterize the uncertainty of the model in evaluating the assembly process.  ...  The component model is then developed by combining the assembly process structure with the latent defect caused by the assembly process.  ... 
doi:10.3390/e20030164 pmid:33265255 fatcat:vigpg2gwy5gbzmbvnl7jhzqv6u

Improving auto-encoder novelty detection using channel attention and entropy minimization [article]

Miao Tian, Dongyan Guo, Ying Cui, Xiang Pan, Shengyong Chen
2021 arXiv   pre-print
Secondly, we apply the information entropy into the latent layer to make it sparse and constrain the expression of diversity.  ...  Auto-encoder is often used for novelty detection.  ...  Method AUC Generator and discriminator 0.619 With latent entropy loss 0.642 With channel attention 0.655  ... 
arXiv:2007.01682v2 fatcat:uaqdustvnfdcjcumuzbac2yh4q

Dense Uncertainty Estimation via an Ensemble-based Conditional Latent Variable Model [article]

Jing Zhang, Yuchao Dai, Mehrtash Harandi, Yiran Zhong, Nick Barnes, Richard Hartley
2021 arXiv   pre-print
In current aleatoric uncertainty estimation frameworks, it is often neglected that the aleatoric uncertainty is an inherent attribute of the data and can only be correctly estimated with an unbiased oracle  ...  For epistemic uncertainty estimation, we argue that the internal variable in a conditional latent variable model is another source of epistemic uncertainty to model the predictive distribution and explore  ...  Predictions and the generated uncertainty maps for camouflaged object detection.  ... 
arXiv:2111.11055v1 fatcat:lncxuj3qcfd6pnbd6j3djh3o4y

The Hidden Uncertainty in a Neural Networks Activations [article]

Janis Postels, Hermann Blum, Yannick Strümpler, Cesar Cadena, Roland Siegwart, Luc Van Gool, Federico Tombari
2021 arXiv   pre-print
The distribution of a neural network's latent representations has been successfully used to detect out-of-distribution (OOD) data.  ...  We first empirically verify that epistemic uncertainty can be identified with the surprise, thus the negative log-likelihood, of observing a particular latent representation.  ...  The aleatoric uncertainty estimated with the conditional entropy can detect misclassified pixel with high AUROC, matching the performance of the softmax entropy.  ... 
arXiv:2012.03082v2 fatcat:2txmq45dhbaytgoc7vn6dyvwuu

Edge Detection Method for Latent Fingerprint Images Using Intuitionistic Type-2 Fuzzy Entropy

Devarasan Ezhilmaran, Manickam Adhiyaman
2016 Cybernetics and Information Technologies  
In this work, a new distance measure has been proposed for latent fingerprint edge detection using Intuitionistic Type-2 Fuzzy Entropy (IT2FE) and a comprehensible definition is made for Intuitionistic  ...  The edge detection is carried out with the proposed method and the results are discovered better regarding existing method.  ...  In this proposed work, a new distance measure has been used with respect to Intuitionistic Type-2 Fuzzy Entropy (IT2FE) for latent fingerprint edge detection (see Fig. 2 ).  ... 
doi:10.1515/cait-2016-0044 fatcat:5xtjhofiz5glnmzbdlc7fzm7ma

Deep Bayesian Gaussian Processes for Uncertainty Estimation in Electronic Health Records [article]

Yikuan Li, Shishir Rao, Abdelaali Hassaine, Rema Ramakrishnan, Yajie Zhu, Dexter Canoy, Gholamreza Salimi-Khorshidi, Thomas Lukasiewicz, Kazem Rahimi
2020 arXiv   pre-print
from the higher level latent space.  ...  In this paper, we merge features of the deep Bayesian learning framework with deep kernel learning to leverage the strengths of both methods for more comprehensive uncertainty estimation.  ...  In this section, we analysed the uncertainty of Bayesian embeddings by measuring its entropy for the DBGP model to understand how the embedding affects the uncertainty in the latent representation.  ... 
arXiv:2003.10170v1 fatcat:6awlh6enknd2pd6gjnsa4mmg4a

Modeling class cohesion as mixtures of latent topics

Yixun Liu, Denys Poshyvanyk, Rudolf Ferenc, Tibor Gyimothy, Nikos Chrisochoides
2009 2009 IEEE International Conference on Software Maintenance  
The measure, named as Maximal Weighted Entropy, utilizes the Latent Dirichlet Allocation technique and information entropy measures to quantitatively evaluate the cohesion of classes in software.  ...  Maximal Weighted Entropy.  ...  We would like to thank Zheng "Eddy" Zhang for her help with LDA tool, Tibor Bakota for verifying the statistical results and comments. This  ... 
doi:10.1109/icsm.2009.5306318 dblp:conf/icsm/LiuPFGC09 fatcat:rfc6bl6h75au7pamlowd3sbaqq
« Previous Showing results 1 — 15 out of 7,662 results