134 Hits in 5.8 sec

Training Neural Networks for Likelihood/Density Ratio Estimation [article]

George V. Moustakides, Kalliopi Basioti
2019 arXiv   pre-print
For most well known problems in Detection and Hypothesis testing we develop solutions by providing neural network based estimates of the likelihood ratio or its transformations.  ...  This task necessitates the definition of proper optimizations which can be used for the training of the network.  ...  u 1 (X, θ 1 ) is a neural network estimate of the log-likelihood ratio of a single sample.  ... 
arXiv:1911.00405v2 fatcat:hvaui6cntfhglpwhocvnpqvgv4

Unified Perspective on Probability Divergence via Maximum Likelihood Density Ratio Estimation: Bridging KL-Divergence and Integral Probability Metrics [article]

Masahiro Kato and Masaaki Imaizumi and Kentaro Minami
2022 arXiv   pre-print
This paper provides a unified perspective for the Kullback-Leibler (KL)-divergence and the integral probability metrics (IPMs) from the perspective of maximum likelihood density-ratio estimation (DRE).  ...  To develop the estimation problem, we construct an unconstrained maximum likelihood estimator to perform DRE with a stratified sampling scheme.  ...  Definition D.5 (ReLU neural networks; Schmidt-Hieber, 2020).  ... 
arXiv:2201.13127v1 fatcat:sps2pjdpyveotmz3k2x3blg3vy

Sequential Density Ratio Estimation for Simultaneous Optimization of Speed and Accuracy [article]

Akinori F. Ebihara, Taiki Miyagawa, Kazuyuki Sakurai, Hitoshi Imaoka
2021 arXiv   pre-print
The SPRT-TANDEM sequentially estimates the log-likelihood ratio of two alternative hypotheses by leveraging a novel Loss function for Log-Likelihood Ratio estimation (LLLR) while allowing correlations  ...  Here, we propose the SPRT-TANDEM, a deep neural network-based SPRT algorithm that overcomes the above two obstacles.  ...  Acknowledgements The authors thank anonymous reviewers for their careful reading to improve the manuscript. We would also like to thank Hirofumi Nakayama and Yuka Fujii for insightful discussions.  ... 
arXiv:2006.05587v3 fatcat:3ajvwnrhy5atlmix2lgqe24qge

Estimation of spatially varying heat transfer coefficient from a flat plate with flush mounted heat sources using Bayesian inference

Pradeep S Jakkareddy, C. Balaji
2016 Journal of Physics, Conference Series  
To speed up the estimation, the forward model is replaced by an artificial neural network. The mean, maximum-a-posteriori and standard deviation of the estimated parameters 'a' and 'b' are reported.  ...  input to a computationally less complex problem of conjugate conduction in the flat plate (15mm thickness) and temperature distributions at the bottom of the plate which is a more convenient location for  ...  |x i+1 −j ) p(x i j |x i+1 −j ) is called as likelihood density ratio(with uniform prior) or PPDF density ratio (with normal prior) and can be calculated from Eqns.19 and 20 The ratio q(x i j |x * j ,x  ... 
doi:10.1088/1742-6596/745/3/032094 fatcat:bpx6bgndijcx7ju7hryknpxyhy

An Introduction to Face Recognition Technology

Shang-Hung Lin
2000 Informing Science  
Several famous face recognition algorithms, such as eigenfaces and neural networks, will also be explained.  ...  For the applications of videophone and teleconferencing, the assistance of face recognition also provides a more efficient coding scheme.  ...  The neural network approach, though some variants of the algorithm work on feature extraction as well, mainly provides sophisticated modeling scheme for estimating likelihood densities in the pattern recognition  ... 
doi:10.28945/569 fatcat:k7ww3kquljdkbkjpfc6pthoqly

Random Finite Set-Based Anomaly Detection for Safety Monitoring in Construction Sites

Ammar Kamoona, Amirali Khodadadian Gostar, Ruwan Tennakoon, Alireza Bab-Hadiashar, David Accadia, Joshua Thorpe, Reza Hoseinnezhad
2019 IEEE Access  
The proposed system is comprised of three steps: 1) applying a deep neural network to extract people in the image; 2) extracting particularly engineered features from each blob returned by the deep neural  ...  Low visibility hazard detection in construction sites is a crucial task for prevention of fatal accidents.  ...  The proposed solution is comprised of three steps: (i) using a deep neural network for people detection, (ii) extracting features from each blob that is returned by the neural network, (iii) applying anomaly  ... 
doi:10.1109/access.2019.2932137 fatcat:grnmmdtmsrf73iteds73z2cnoy

Reducing the Variance of Variational Estimates of Mutual Information by Limiting the Critic's Hypothesis Space to RKHS [article]

P Aditya Sreekar, Ujjwal Tiwari, Anoop Namboodiri
2020 arXiv   pre-print
Recent methods realize parametric probability distributions or critic as a neural network to approximate unknown density ratios.  ...  The approximated density ratios are used to estimate different variational lower bounds of MI.  ...  In these methods, a critic parameterized as a neural network is trained to approximate unknown density ratios.  ... 
arXiv:2011.08651v1 fatcat:6b5u4bqyo5erzcdrfq475wytnq

Mining gold from implicit models to improve likelihood-free inference [article]

Johann Brehmer, Gilles Louppe, Juan Pavez, Kyle Cranmer
2019 arXiv   pre-print
We show that additional information, such as the joint likelihood ratio and the joint score, can often be extracted from simulators and used to augment the training data for these surrogate models.  ...  networks.  ...  We are grateful to Jan-Matthis Lückmann for helping us automate the calculation of the joint likelihood ratio and joint score in PYRO and to all participants of the Likelihood-free inference workshop at  ... 
arXiv:1805.12244v4 fatcat:ttk5fkjwdbghtiqd7tfpisr4ii

Learning Uncertainties the Frequentist Way: Calibration and Correlation in High Energy Physics [article]

Rikab Gambhir, Benjamin Nachman, Jesse Thaler
2022 arXiv   pre-print
In this paper, we present a machine learning framework for performing frequentist maximum likelihood inference with Gaussian uncertainty estimation, which also quantifies the mutual information between  ...  the Kullback-Leibler divergence -- parametrized with a novel Gaussian Ansatz -- to enable a simultaneous extraction of the maximum likelihood values, uncertainties, and mutual information in a single training  ...  , we can use T as an estimate of the log-likelihood density ratio.  ... 
arXiv:2205.03413v2 fatcat:p7fxy43zyjezhdgqqprtlmph3e

High-dimensional Metric Combining for Non-coherent Molecular Signal Detection [article]

Zhuangkun Wei, Weisi Guo, Bin Li, Jerome Charmet, Chenglin Zhao
2019 arXiv   pre-print
Then, we design a generalised blind detection algorithm that utilizes the Parzen approximation and its probabilistic neural network (Parzen-PNN) to detect information bits.  ...  If the channel is unknown, we cannot easily achieve traditional coherent channel estimation and cancellation, and the impact of ISI will be more severe.  ...  blind detection, we suggest the Parzen technique with a probabilistic neural network (Parzen-PNN) to approximate the likelihood density, and further detect the information bits.  ... 
arXiv:1901.11422v1 fatcat:ypyvek3pknhjddngzyp6sb46ca

Enhanced Deep Learning Assisted Convolutional Neural Network for Heart Disease Prediction on the Internet of Medical Things Platform

Yuanyuan Pan, Minghuan Fu, Biao Cheng, Xuefei Tao, Jing Guo
2020 IEEE Access  
Thus an exponential formulation is used to describe the likelihood density for cardiovascular events of a patient at an elapsed duration t after release: ( | ; , ) = exp(− ) = exp{−( + ) } = {−(∑ + =1  ...  (iii) Efficiency Ratio determination The deep convolutional neural network (or diagnostic) model's efficiency quality depends heavily on the DNN model classification while the training process.  ... 
doi:10.1109/access.2020.3026214 fatcat:a6uzyx7hwrhsphgt3qvi53bkbe

Sequential Monte Carlo Methods to Train Neural Network Models

J. F. G. de Freitas, M. Niranjan, A. H. Gee, A. Doucet
2000 Neural Computation  
We discuss a novel strategy for training neural networks using sequential Monte Carlo algorithms and propose a new hybrid gradient descent/ sampling importance resampling algorithm (HySIR).  ...  In this context, we are able to estimate the one-step-ahead probability density functions of the options prices. Neural Computation 12, 955-993 (2000)  ...  We are also very grateful to the referees for their valuable comments. J. F. G.  ... 
doi:10.1162/089976600300015664 pmid:10770839 fatcat:jzytx4cwyzewlmdjfv5p2bvrya

Lane Change Prediction Using Gaussian Classification, Support Vector Classification and Neural Network Classifiers

Olivér Rákos, Szilárd Aradi, Tamás Bécsi
2020 Periodica Polytechnica Transportation Engineering  
It has been investigated whether it is possible to reliably classify lane-changing maneuvers in a highway situation using learning algorithms such as Gaussian-classifier, SVM, and LSTM neural networks.  ...  Different strategies for labeling the input sequences were tested.  ...  Approximately the same result were achieved for the SVM trained by the time series of ∆x k ( ) values and neural network classifiers trained by longitudinal values ∆x k v k a k ( ) ( ) ( ) , , .  ... 
doi:10.3311/pptr.15849 fatcat:ghp253arsbacphskf7fexr4g3m

Identification of Vehicle Dynamics Parameters Using Simulation-based Inference [article]

Ali Boyali, Simon Thompson, David Robert Wong
2021 arXiv   pre-print
We demonstrate in this paper that it can handle the identification of highly nonlinear vehicle dynamics parameters and gives accurate estimates of the parameters for the governing equations.  ...  Identifying tire and vehicle parameters is an essential step in designing control and planning algorithms for autonomous vehicles.  ...  ACKNOWLEDGMENT This work was presented at the workshop Autoware -ROS-based OSS for Autonomous Driving WS26 IV2021.  ... 
arXiv:2108.12114v1 fatcat:5lh23uh7qzaozek5h7uf54hjly

CLUB: A Contrastive Log-ratio Upper Bound of Mutual Information [article]

Pengyu Cheng, Weituo Hao, Shuyang Dai, Jiachang Liu, Zhe Gan, Lawrence Carin
2020 arXiv   pre-print
Based on this upper bound, we introduce a MI minimization training scheme and further accelerate it with a negative sampling strategy.  ...  Simulation studies on Gaussian distributions show the reliable estimation ability of CLUB.  ...  Acknowledgements Thanks to Dongruo Zhou from UCLA for helpful discussions on network expressiveness.  ... 
arXiv:2006.12013v6 fatcat:ilsrj3qxsjerfgkqr5c7sq4nuq
« Previous Showing results 1 — 15 out of 134 results