37,911 Hits in 6.9 sec

On Recovering from Modeling Errors Using Testing Bayesian Networks

Haiying Huang, Adnan Darwiche
2021 International Conference on Machine Learning  
We then identify a class of Bayesian Networks and queries which allow one to fully recover from such modeling errors if one can choose Conditional Probability Tables (CPTs) dynamically based on evidence  ...  Finally, we present empirical results that illustrate the promise of TBNs as a tool for recovering from certain modeling errors in the context of supervised learning.  ...  Acknowledgements This work has been partially supported by grants from NSF IIS-1910317, ONR N00014-18-1-2561, and DARPA N66001-17-2-4032.  ... 
dblp:conf/icml/HuangD21 fatcat:zjvqvt4ko5gyzblxjcimqjkgoi

Bayesian Neural Ordinary Differential Equations [article]

Raj Dandekar, Karen Chung, Vaibhav Dixit, Mohamed Tarek, Aslan Garcia-Valadez, Krishna Vishal Vemula, Chris Rackauckas
2022 arXiv   pre-print
On the MNIST dataset, we achieve a posterior sample accuracy of 98.5% on the test ensemble of 10,000 images.  ...  We demonstrate the successful integration of Neural ODEs with the above Bayesian inference frameworks on classical physical systems, as well as on standard machine learning datasets like MNIST, using GPU  ...  test error test error RK-Net 0.47 % Chen (2018) ODE-Net 0.42 % Chen (2018) Bayesian Alex-Net 1 % Shridhar (2019) Bayesian LeNet-5 2 % Shridhar (2019) Bayesian Neural ODE (ensemble) 0.78 % Our study adapative  ... 
arXiv:2012.07244v4 fatcat:rmlwv5ywknch3mignxbc5hciwq

Some Variations on the PC Algorithm

Joaquín Abellán, Manuel Gómez-Olmedo, Serafín Moral
2006 European Workshop on Probabilistic Graphical Models  
test, to study the refinement of the learned network by a greedy optimization of a Bayesian score, and to solve link ambiguities taking into account a measure of their strength.  ...  This paper proposes some possible modifications on the PC basic learning algorithm and makes some experiments to study their behaviour.  ...  Then, we have tried to recover the original network from the samples by using the different variations of the PC algorithm including the orientation step.  ... 
dblp:conf/pgm/AbellanGM06 fatcat:gtnkmfzrpnf6xjhw7sxubfo7wy

Modelling galaxy emission-line kinematics using self-supervised learning

James Dawson
2021 Zenodo  
In this talk I will present the network's ability to model the kinematics of cold gas in galaxies with an emphasis on recovering physical parameters and accompanying modelling errors.  ...  In order to assist in the timely exploitation of these vast datasets we have explored the use of self-supervised, physics aware neural networks capable of Bayesian kinematic modelling of galaxies.  ...  Six galax- ies drawn from the WISDOM project, with high 'spatial resolution', were trained and tested on using our model.  ... 
doi:10.5281/zenodo.5607462 fatcat:j7ocrqvhobb3vc3yedhhvdv3hm

Bayesian 3D Human Motion Capture Using Factored Particle Filtering

Abdallah Dib, Cedric Rose, Francois Charpillet
2010 2010 22nd IEEE International Conference on Tools with Artificial Intelligence  
The system uses a dynamic bayesian network and a factored particle filtering algorithm.  ...  The capacity of the system to recover after occlusion by obstacles was tested on simulated movements in a virtual scene.  ...  Finally, we tested the capacity of the system to recover from occluded positions.  ... 
doi:10.1109/ictai.2010.131 dblp:conf/ictai/DibRC10 fatcat:b4x6jiy73fedvd64vr7xcmaf7i

On some fundamental challenges in monitoring epidemics

Vaiva Vasiliauskaite, Nino Antulov-Fantulin, Dirk Helbing
2021 Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences  
Nevertheless, some of the errors can be corrected for, using scientific knowledge of the spreading dynamics and the measurement process.  ...  or recovered).  ...  We thank Lucas Böttcher for inspiring discussions and scientific exchange in the initial phase of the project, which related to models of statistical measurement as presented in [20] and to empirical  ... 
doi:10.1098/rsta.2021.0117 pmid:34802270 pmcid:PMC8607144 fatcat:jmvr3kjre5dq5mddpdr3t735uq

Introspective Visuomotor Control: Exploiting Uncertainty in Deep Visuomotor Control for Failure Recovery [article]

Chia-Man Hung, Li Sun, Yizhe Wu, Ioannis Havoutis, Ingmar Posner
2021 arXiv   pre-print
However, imitation learning-based visuomotor control approaches tend to suffer from a common limitation, lacking the ability to recover from an out-of-distribution state caused by compounding errors.  ...  In this paper, instead of using tactile feedback or explicitly detecting the failure through vision, we investigate using the uncertainty of a policy neural network.  ...  We also thank Ruoqi He, Hala Lamdouar, Walter Goodwin and Oliver Groth for proofreading and useful discussions, and the reviewers for valuable feedback.  ... 
arXiv:2103.11881v1 fatcat:nozep6rr7bb7fi7ghredawqyuq

Data-driven prediction and origin identification of epidemics in population networks

Karen Larson, Georgios Arampatzis, Clark Bowman, Zhizhong Chen, Panagiotis Hadjidoukas, Costas Papadimitriou, Petros Koumoutsakos, Anastasios Matzavinos
2021 Royal Society Open Science  
We introduce a Bayesian uncertainty quantification framework to infer model parameters for a disease spreading on a network of communities from limited, noisy observations; the state-of-the-art computational  ...  Effective intervention strategies for epidemics rely on the identification of their origin and on the robustness of the predictions made by network disease models.  ...  Parts of this research were conducted using computational resources and services at the Center for Computation and Visualization, Brown University.  ... 
doi:10.1098/rsos.200531 pmid:33614060 pmcid:PMC7890494 fatcat:eklj3h2kond3bfguse6qf7mx5q

Introduction to the vol. 47, no. 1, 2020

Maomi Ueno
2020 Behaviormetrika  
categorical variables using the Fisher kernel for Bayesian networks.  ...  This study also experimentally demonstrates how this kernel can be used to find subsets of observations that we see as representative for the underlying Bayesian network model.  ... 
doi:10.1007/s41237-020-00106-8 fatcat:atdtrfmjxndktolwnpkzjzvxii

How Good Is Crude MDL for Solving the Bias-Variance Dilemma? An Empirical Investigation Based on Bayesian Networks

Nicandro Cruz-Ramírez, Héctor Gabriel Acosta-Mesa, Efrén Mezura-Montes, Alejandro Guerra-Hernández, Guillermo de Jesús Hoyos-Rivera, Rocío Erandi Barrientos-Martínez, Karina Gutiérrez-Fragoso, Luis Alonso Nava-Fernández, Patricia González-Gaspar, Elva María Novoa-del-Toro, Vicente Josué Aguilera-Rueda, María Yaneli Ameca-Alducin (+1 others)
2014 PLoS ONE  
We carry out these experiments using a specific model: a Bayesian network.  ...  not necessarily need be the gold-standard ones.  ...  8, 16, 32, 64 and 128 are generated) and different Bayesian network structures: the one with the best MDL score, the complete, the independent, the maximum error, the minimum error and the Chow-Liu networks  ... 
doi:10.1371/journal.pone.0092866 pmid:24671204 pmcid:PMC3966834 fatcat:zuyr7qh5qzapfdhsfd2szm5i5q

Unsupervised body scheme learning through self-perception

Jurgen Sturm, Christian Plagemann, Wolfram Burgard
2008 2008 IEEE International Conference on Robotics and Automation  
From this repertoire of local models, we construct a Bayesian network for the full system using the pose prediction accuracy on a separate cross validation data set as the criterion for model selection  ...  Our approach yields a compact Bayesian network for the robot's kinematic structure including the forward and inverse models relating action signals and body pose.  ...  Left: Screen-shot from the simulated robot. Middle: One of the potential Bayesian networks after the first training sample: the correct kinematic chain is not yet recovered.  ... 
doi:10.1109/robot.2008.4543718 dblp:conf/icra/SturmPB08 fatcat:fhdbqqzgn5ayrahoewcd7k5hla

Simultaneous Measurement Imputation and Outcome Prediction for Achilles Tendon Rupture Rehabilitation [article]

Charles Hamesse, Ruibo Tu, Paul Ackermann, Hedvig Kjellström, Cheng Zhang
2019 arXiv   pre-print
We evaluate our model on a real-life ATR clinical cohort, comparing with various baselines.  ...  However, it is challenging to train an automatic method for predicting the ATR rehabilitation outcome from treatment data, due to a massive amount of missing entries in the data recorded from ATR patients  ...  Our proposed model with Bayesian neural network is used for this evaluation.  ... 
arXiv:1810.03435v2 fatcat:pvevptpqrzf3phckyeuabgwvue

OutbreakFlow: Model-based Bayesian inference of disease outbreak dynamics with invertible neural networks and its application to the COVID-19 pandemics in Germany [article]

Stefan T. Radev, Frederik Graw, Simiao Chen, Nico T. Mutters, Vanessa M. Eichel, Till Bärnighausen, Ullrich Köthe
2021 arXiv   pre-print
Apart from their scientific merit, these models are often used to inform political decisions and intervention measures during an ongoing outbreak.  ...  which have to be repeated from scratch for every application of a given model.  ...  The model was trained on a time-period from March 01 until May 21 using wide prior distributions across plausible parameter ranges from previous literature [8, 22] Our model was able to recover the observed  ... 
arXiv:2010.00300v4 fatcat:jqjs5cgdwbe3tnsfttrqtutiaq

Non-linear carbon dioxide determination using infrared gas sensors and neural networks with Bayesian regularization

King-Tong Lau, Weimin Guo, Breda Kiernan, Conor Slater, Dermot Diamond
2009 Sensors and actuators. B, Chemical  
Carbon dioxide gas concentration determination using infrared gas sensors combined with Bayesian regularizing neural networks is presented in this work.  ...  The Bayesian strategy was used to regularize the training of the back propagation neural network with a Levenberg-Marquardt (LM) algorithm.  ...  Neural networks were also well used in environmental applications from LFG production models to calibration of sensors [17] .  ... 
doi:10.1016/j.snb.2008.11.030 fatcat:ybrkjczevjbl7ftf7i5npsoc2u

High-dimensional Bayesian network inference from systems genetics data using genetic node ordering [article]

Lingfei Wang, Pieter Audenaert, Tom Michoel
2018 bioRxiv   pre-print
Bayesian networks provide an elegant statistical approach for multi-trait genetic mapping and modelling causal trait relationships.  ...  However, inferring Bayesian gene networks from high-dimensional genetics and genomics data is challenging, because the number of possible networks scales super-exponentially with the number of nodes, and  ...  Lasso-findr and lasso-random Bayesian networks using penalized regression on ordered nodes To infer a more refined sparse Bayesian network from a maximal DAG, we performed hypothesis testing for every  ... 
doi:10.1101/501460 fatcat:aj4rx7vfffhrhh2njdfqfipt3u
« Previous Showing results 1 — 15 out of 37,911 results