The Experts below are selected from a list of 7857 Experts worldwide ranked by ideXlab platform
Lawrence Carin - One of the best experts on this subject based on the ideXlab platform.
-
a convergence analysis for a class of practical Variance Reduction stochastic gradient mcmc
Science in China Series F: Information Sciences, 2019Co-Authors: Changyou Chen, Wenlin Wang, Yizhe Zhang, Lawrence CarinAbstract:Stochastic gradient Markov chain Monte Carlo (SG-MCMC) has been developed as a flexible family of scalable Bayesian sampling algorithms. However, there has been little theoretical analysis of the impact of minibatch size to the algorithm’s convergence rate. In this paper, we prove that at the beginning of an SG-MCMC algorithm, i.e., under limited computational budget/time, a larger minibatch size leads to a faster decrease of the mean squared error bound. The reason for this is due to the prominent noise in small minibatches when calculating stochastic gradients, motivating the necessity of Variance Reduction in SG-MCMC for practical use. By borrowing ideas from stochastic optimization, we propose a simple and practical Variance-Reduction Technique for SG-MCMC, that is efficient in both computation and storage. More importantly, we develop the theory to prove that our algorithm induces a faster convergence rate than standard SG-MCMC. A number of large-scale experiments, ranging from Bayesian learning of logistic regression to deep neural networks, validate the theory and demonstrate the superiority of the proposed Variance-Reduction SG-MCMC framework.
-
a convergence analysis for a class of practical Variance Reduction stochastic gradient mcmc
arXiv: Machine Learning, 2017Co-Authors: Changyou Chen, Wenlin Wang, Yizhe Zhang, Lawrence CarinAbstract:Stochastic gradient Markov Chain Monte Carlo (SG-MCMC) has been developed as a flexible family of scalable Bayesian sampling algorithms. However, there has been little theoretical analysis of the impact of minibatch size to the algorithm's convergence rate. In this paper, we prove that under a limited computational budget/time, a larger minibatch size leads to a faster decrease of the mean squared error bound (thus the fastest one corresponds to using full gradients), which motivates the necessity of Variance Reduction in SG-MCMC. Consequently, by borrowing ideas from stochastic optimization, we propose a practical Variance-Reduction Technique for SG-MCMC, that is efficient in both computation and storage. We develop theory to prove that our algorithm induces a faster convergence rate than standard SG-MCMC. A number of large-scale experiments, ranging from Bayesian learning of logistic regression to deep neural networks, validate the theory and demonstrate the superiority of the proposed Variance-Reduction SG-MCMC framework.
N Freud - One of the best experts on this subject based on the ideXlab platform.
-
monte carlo simulation of prompt γ ray emission in proton therapy using a specific track length estimator
Physics in Medicine and Biology, 2015Co-Authors: El W Kanawati, J M Letang, D Dauvergne, M Pinto, David Sarrut, E Testa, N FreudAbstract:A Monte Carlo (MC) Variance Reduction Technique is developed for prompt-γ emitters calculations in proton therapy. Prompt-γ emitted through nuclear fragmentation reactions and exiting the patient during proton therapy could play an important role to help monitoring the treatment. However, the estimation of the number and the energy of emitted prompt-γ per primary proton with MC simulations is a slow process. In order to estimate the local distribution of prompt-γ emission in a volume of interest for a given proton beam of the treatment plan, a MC Variance Reduction Technique based on a specific track length estimator (TLE) has been developed. First an elemental database of prompt-γ emission spectra is established in the clinical energy range of incident protons for all elements in the composition of human tissues. This database of the prompt-γ spectra is built offline with high statistics. Regarding the implementation of the prompt-γ TLE MC tally, each proton deposits along its track the expectation of the prompt-γ spectra from the database according to the proton kinetic energy and the local material composition. A detailed statistical study shows that the relative efficiency mainly depends on the geometrical distribution of the track length. Benchmarking of the proposed prompt-γ TLE MC Technique with respect to an analogous MC Technique is carried out. A large relative efficiency gain is reported, ca. 105.
-
monte carlo simulation of prompt γ ray emission in proton therapy using a specific track length estimator
Physics in Medicine and Biology, 2015Co-Authors: El W Kanawati, J M Letang, D Dauvergne, M Pinto, David Sarrut, E Testa, N FreudAbstract:A Monte Carlo (MC) Variance Reduction Technique is developed for prompt-γ emitters calculations in proton therapy. Prompt-γ emitted through nuclear fragmentation reactions and exiting the patient during proton therapy could play an important role to help monitoring the treatment. However, the estimation of the number and the energy of emitted prompt-γ per primary proton with MC simulations is a slow process. In order to estimate the local distribution of prompt-γ emission in a volume of interest for a given proton beam of the treatment plan, a MC Variance Reduction Technique based on a specific track length estimator (TLE) has been developed. First an elemental database of prompt-γ emission spectra is established in the clinical energy range of incident protons for all elements in the composition of human tissues. This database of the prompt-γ spectra is built offline with high statistics. Regarding the implementation of the prompt-γ TLE MC tally, each proton deposits along its track the expectation of the prompt-γ spectra from the database according to the proton kinetic energy and the local material composition. A detailed statistical study shows that the relative efficiency mainly depends on the geometrical distribution of the track length. Benchmarking of the proposed prompt-γ TLE MC Technique with respect to an analogous MC Technique is carried out. A large relative efficiency gain is reported, ca. 105.
Changyou Chen - One of the best experts on this subject based on the ideXlab platform.
-
a convergence analysis for a class of practical Variance Reduction stochastic gradient mcmc
Science in China Series F: Information Sciences, 2019Co-Authors: Changyou Chen, Wenlin Wang, Yizhe Zhang, Lawrence CarinAbstract:Stochastic gradient Markov chain Monte Carlo (SG-MCMC) has been developed as a flexible family of scalable Bayesian sampling algorithms. However, there has been little theoretical analysis of the impact of minibatch size to the algorithm’s convergence rate. In this paper, we prove that at the beginning of an SG-MCMC algorithm, i.e., under limited computational budget/time, a larger minibatch size leads to a faster decrease of the mean squared error bound. The reason for this is due to the prominent noise in small minibatches when calculating stochastic gradients, motivating the necessity of Variance Reduction in SG-MCMC for practical use. By borrowing ideas from stochastic optimization, we propose a simple and practical Variance-Reduction Technique for SG-MCMC, that is efficient in both computation and storage. More importantly, we develop the theory to prove that our algorithm induces a faster convergence rate than standard SG-MCMC. A number of large-scale experiments, ranging from Bayesian learning of logistic regression to deep neural networks, validate the theory and demonstrate the superiority of the proposed Variance-Reduction SG-MCMC framework.
-
a convergence analysis for a class of practical Variance Reduction stochastic gradient mcmc
arXiv: Machine Learning, 2017Co-Authors: Changyou Chen, Wenlin Wang, Yizhe Zhang, Lawrence CarinAbstract:Stochastic gradient Markov Chain Monte Carlo (SG-MCMC) has been developed as a flexible family of scalable Bayesian sampling algorithms. However, there has been little theoretical analysis of the impact of minibatch size to the algorithm's convergence rate. In this paper, we prove that under a limited computational budget/time, a larger minibatch size leads to a faster decrease of the mean squared error bound (thus the fastest one corresponds to using full gradients), which motivates the necessity of Variance Reduction in SG-MCMC. Consequently, by borrowing ideas from stochastic optimization, we propose a practical Variance-Reduction Technique for SG-MCMC, that is efficient in both computation and storage. We develop theory to prove that our algorithm induces a faster convergence rate than standard SG-MCMC. A number of large-scale experiments, ranging from Bayesian learning of logistic regression to deep neural networks, validate the theory and demonstrate the superiority of the proposed Variance-Reduction SG-MCMC framework.
J M Letang - One of the best experts on this subject based on the ideXlab platform.
-
neutron track length estimator for gate monte carlo dose calculation in radiotherapy
Physics in Medicine and Biology, 2018Co-Authors: H Elazhar, J M Letang, T Deschler, A Nourreddine, Nicolas ArborAbstract:: The out-of-field dose in radiation therapy is a growing concern in regards to the late side-effects and secondary cancer induction. In high-energy x-ray therapy, the secondary neutrons generated through photonuclear reactions in the accelerator are part of this secondary dose. The neutron dose is currently not estimated by the treatment planning system while it appears to be preponderant for distances greater than 50 cm from the isocenter. Monte Carlo simulation has become the gold standard for accurately calculating the neutron dose under specific treatment conditions but the method is also known for having a slow statistical convergence, which makes it difficult to be used on a clinical basis. The neutron track length estimator, a neutron Variance Reduction Technique inspired by the track length estimator method has thus been developped for the first time in the Monte Carlo code GATE to allow a fast computation of the neutron dose in radiotherapy. The details of its implementation, as well as the comparison of its performances against the analog MC method, are presented here. A gain of time from 15 to 400 can be obtained by our method, with a mean difference in the dose calculation of about 1% in comparison with the analog MC method.
-
monte carlo simulation of prompt γ ray emission in proton therapy using a specific track length estimator
Physics in Medicine and Biology, 2015Co-Authors: El W Kanawati, J M Letang, D Dauvergne, M Pinto, David Sarrut, E Testa, N FreudAbstract:A Monte Carlo (MC) Variance Reduction Technique is developed for prompt-γ emitters calculations in proton therapy. Prompt-γ emitted through nuclear fragmentation reactions and exiting the patient during proton therapy could play an important role to help monitoring the treatment. However, the estimation of the number and the energy of emitted prompt-γ per primary proton with MC simulations is a slow process. In order to estimate the local distribution of prompt-γ emission in a volume of interest for a given proton beam of the treatment plan, a MC Variance Reduction Technique based on a specific track length estimator (TLE) has been developed. First an elemental database of prompt-γ emission spectra is established in the clinical energy range of incident protons for all elements in the composition of human tissues. This database of the prompt-γ spectra is built offline with high statistics. Regarding the implementation of the prompt-γ TLE MC tally, each proton deposits along its track the expectation of the prompt-γ spectra from the database according to the proton kinetic energy and the local material composition. A detailed statistical study shows that the relative efficiency mainly depends on the geometrical distribution of the track length. Benchmarking of the proposed prompt-γ TLE MC Technique with respect to an analogous MC Technique is carried out. A large relative efficiency gain is reported, ca. 105.
-
monte carlo simulation of prompt γ ray emission in proton therapy using a specific track length estimator
Physics in Medicine and Biology, 2015Co-Authors: El W Kanawati, J M Letang, D Dauvergne, M Pinto, David Sarrut, E Testa, N FreudAbstract:A Monte Carlo (MC) Variance Reduction Technique is developed for prompt-γ emitters calculations in proton therapy. Prompt-γ emitted through nuclear fragmentation reactions and exiting the patient during proton therapy could play an important role to help monitoring the treatment. However, the estimation of the number and the energy of emitted prompt-γ per primary proton with MC simulations is a slow process. In order to estimate the local distribution of prompt-γ emission in a volume of interest for a given proton beam of the treatment plan, a MC Variance Reduction Technique based on a specific track length estimator (TLE) has been developed. First an elemental database of prompt-γ emission spectra is established in the clinical energy range of incident protons for all elements in the composition of human tissues. This database of the prompt-γ spectra is built offline with high statistics. Regarding the implementation of the prompt-γ TLE MC tally, each proton deposits along its track the expectation of the prompt-γ spectra from the database according to the proton kinetic energy and the local material composition. A detailed statistical study shows that the relative efficiency mainly depends on the geometrical distribution of the track length. Benchmarking of the proposed prompt-γ TLE MC Technique with respect to an analogous MC Technique is carried out. A large relative efficiency gain is reported, ca. 105.
El W Kanawati - One of the best experts on this subject based on the ideXlab platform.
-
monte carlo simulation of prompt γ ray emission in proton therapy using a specific track length estimator
Physics in Medicine and Biology, 2015Co-Authors: El W Kanawati, J M Letang, D Dauvergne, M Pinto, David Sarrut, E Testa, N FreudAbstract:A Monte Carlo (MC) Variance Reduction Technique is developed for prompt-γ emitters calculations in proton therapy. Prompt-γ emitted through nuclear fragmentation reactions and exiting the patient during proton therapy could play an important role to help monitoring the treatment. However, the estimation of the number and the energy of emitted prompt-γ per primary proton with MC simulations is a slow process. In order to estimate the local distribution of prompt-γ emission in a volume of interest for a given proton beam of the treatment plan, a MC Variance Reduction Technique based on a specific track length estimator (TLE) has been developed. First an elemental database of prompt-γ emission spectra is established in the clinical energy range of incident protons for all elements in the composition of human tissues. This database of the prompt-γ spectra is built offline with high statistics. Regarding the implementation of the prompt-γ TLE MC tally, each proton deposits along its track the expectation of the prompt-γ spectra from the database according to the proton kinetic energy and the local material composition. A detailed statistical study shows that the relative efficiency mainly depends on the geometrical distribution of the track length. Benchmarking of the proposed prompt-γ TLE MC Technique with respect to an analogous MC Technique is carried out. A large relative efficiency gain is reported, ca. 105.
-
monte carlo simulation of prompt γ ray emission in proton therapy using a specific track length estimator
Physics in Medicine and Biology, 2015Co-Authors: El W Kanawati, J M Letang, D Dauvergne, M Pinto, David Sarrut, E Testa, N FreudAbstract:A Monte Carlo (MC) Variance Reduction Technique is developed for prompt-γ emitters calculations in proton therapy. Prompt-γ emitted through nuclear fragmentation reactions and exiting the patient during proton therapy could play an important role to help monitoring the treatment. However, the estimation of the number and the energy of emitted prompt-γ per primary proton with MC simulations is a slow process. In order to estimate the local distribution of prompt-γ emission in a volume of interest for a given proton beam of the treatment plan, a MC Variance Reduction Technique based on a specific track length estimator (TLE) has been developed. First an elemental database of prompt-γ emission spectra is established in the clinical energy range of incident protons for all elements in the composition of human tissues. This database of the prompt-γ spectra is built offline with high statistics. Regarding the implementation of the prompt-γ TLE MC tally, each proton deposits along its track the expectation of the prompt-γ spectra from the database according to the proton kinetic energy and the local material composition. A detailed statistical study shows that the relative efficiency mainly depends on the geometrical distribution of the track length. Benchmarking of the proposed prompt-γ TLE MC Technique with respect to an analogous MC Technique is carried out. A large relative efficiency gain is reported, ca. 105.
