-
Bayesian functional linear regression with sparse step functions
Authors:
Paul-Marie Grollemund,
Christophe Abraham,
Meïli Baragatti,
Pierre Pudlo
Abstract:
The functional linear regression model is a common tool to determine the relationship between a scalar outcome and a functional predictor seen as a function of time. This paper focuses on the Bayesian estimation of the support of the coefficient function. To this aim we propose a parsimonious and adaptive decomposition of the coefficient function as a step function, and a model including a prior d…
▽ More
The functional linear regression model is a common tool to determine the relationship between a scalar outcome and a functional predictor seen as a function of time. This paper focuses on the Bayesian estimation of the support of the coefficient function. To this aim we propose a parsimonious and adaptive decomposition of the coefficient function as a step function, and a model including a prior distribution that we name Bayesian functional Linear regression with Sparse Step functions (Bliss). The aim of the method is to recover areas of time which influences the most the outcome. A Bayes estimator of the support is built with a specific loss function, as well as two Bayes estimators of the coefficient function, a first one which is smooth and a second one which is a step function. The performance of the proposed methodology is analysed on various synthetic datasets and is illustrated on a black Périgord truffle dataset to study the influence of rainfall on the production.
△ Less
Submitted 6 January, 2017; v1 submitted 28 April, 2016;
originally announced April 2016.
-
Enhanced Circuit Densities in Epitaxially Defined FinFETs (EDFinFETs) over FinFETs
Authors:
Sushant Mittal,
Aneesh Nainani,
M. C. Abraham,
Saurabh Lodha,
Udayan Ganguly
Abstract:
FinFET technology is prone to suffer from Line Edge Roughness (LER) based VT variation with scaling. To address this, we proposed an Epitaxially Defined (ED) FinFET (EDFinFET) as an alternate to FinFET architecture for 10 nm node and beyond. We showed by statistical simulations that EDFinFET reduces LER based VT variability by 90% and overall variability by 59%. However, EDFinFET consists of wider…
▽ More
FinFET technology is prone to suffer from Line Edge Roughness (LER) based VT variation with scaling. To address this, we proposed an Epitaxially Defined (ED) FinFET (EDFinFET) as an alternate to FinFET architecture for 10 nm node and beyond. We showed by statistical simulations that EDFinFET reduces LER based VT variability by 90% and overall variability by 59%. However, EDFinFET consists of wider fins as the fin widths are not constrained by electrostatics and variability (cf. FinFETs have fin width ~ LG/3 where LG is gate-length). This indicates that EDFinFET based circuits may be less dense. In this study we show that wide fins enable taller fin heights. The ability to engineer multiple STI levels on tall fins enables different transistor widths (i.e. various W/Ls e.g. 1-10) in a single fin. This capability ensures that even though individual EDFinFET devices have ~2x larger footprints than FinFETs, EDFinFET may produce equal or higher circuit density for basic building blocks like inverters or NAND gates for W/Ls of 2 and higher.
△ Less
Submitted 15 April, 2016;
originally announced April 2016.
-
Excursion theory for Brownian motion indexed by the Brownian tree
Authors:
Céline Abraham,
Jean-François Le Gall
Abstract:
We develop an excursion theory for Brownian motion indexed by the Brownian tree, which in many respects is analogous to the classical Itô theory for linear Brownian motion. Each excursion is associated with a connected component of the complement of the zero set of the tree-indexed Brownian motion. Each such connectedcomponent is itself a continuous tree, and we introduce a quantity measuring the…
▽ More
We develop an excursion theory for Brownian motion indexed by the Brownian tree, which in many respects is analogous to the classical Itô theory for linear Brownian motion. Each excursion is associated with a connected component of the complement of the zero set of the tree-indexed Brownian motion. Each such connectedcomponent is itself a continuous tree, and we introduce a quantity measuring the length of its boundary. The collection of boundary lengths coincides with the collection of jumps of a continuous-state branching process with branching mechanism $ψ(u)=\sqrt{8/3}\,u^{3/2}$. Furthermore, conditionally on the boundary lengths, the different excursions are independent, and we determine their conditional distribution in terms of an excursion measure $\mathbb{M}_0$ which is the analog of the Itô measure of Brownian excursions. We provide various descriptions of the excursion measure $\mathbb{M}_0$, and we also determine several explicit distributions, such as the joint distribution of the boundary length and the mass of an excursion under $\mathbb{M}_0$. We use the Brownian snake as a convenient tool for defining and analysing the excursions of our tree-indexed Brownian motion.
△ Less
Submitted 12 September, 2018; v1 submitted 22 September, 2015;
originally announced September 2015.
-
Random maps: proceeding of the Journ{é}es MAS 2014
Authors:
Céline Abraham,
Jérémie Bettinelli,
Gwendal Collet,
Igor Kortchemski
Abstract:
This is a quick survey on some recent works done in the field of random maps.
This is a quick survey on some recent works done in the field of random maps.
△ Less
Submitted 4 December, 2014;
originally announced December 2014.
-
Rescaled bipartite planar maps converge to the Brownian map
Authors:
Céline Abraham
Abstract:
For every integer $n\geq 1$, we consider a random planar map $\mathcal{M}_n$ which is uniformly distributed over the class of all rooted bipartite planar maps with $n$ edges. We prove that the vertex set of $\mathcal{M}_n$ equipped with the graph distance rescaled by the factor $(2n)^{-1/4}$ converges in distribution, in the Gromov-Hausdorff sense, to the Brownian map. This complements several rec…
▽ More
For every integer $n\geq 1$, we consider a random planar map $\mathcal{M}_n$ which is uniformly distributed over the class of all rooted bipartite planar maps with $n$ edges. We prove that the vertex set of $\mathcal{M}_n$ equipped with the graph distance rescaled by the factor $(2n)^{-1/4}$ converges in distribution, in the Gromov-Hausdorff sense, to the Brownian map. This complements several recent results giving the convergence of various classes of random planar maps to the Brownian map.
△ Less
Submitted 23 July, 2014; v1 submitted 20 December, 2013;
originally announced December 2013.
-
Asymptotic global robustness in bayesian decision theory
Authors:
Christophe Abraham,
Benoit Cadre
Abstract:
In Bayesian decision theory, it is known that robustness with respect to the loss and the prior can be improved by adding new observations. In this article we study the rate of robustness improvement with respect to the number of observations n. Three usual measures of posterior global robustness are considered: the (range of the) Bayes actions set derived from a class of loss functions, the max…
▽ More
In Bayesian decision theory, it is known that robustness with respect to the loss and the prior can be improved by adding new observations. In this article we study the rate of robustness improvement with respect to the number of observations n. Three usual measures of posterior global robustness are considered: the (range of the) Bayes actions set derived from a class of loss functions, the maximum regret of using a particular loss when the subjective loss belongs to a given class and the range of the posterior expected loss when the loss function ranges over a class. We show that the rate of convergence of the first measure of robustness is \sqrtn, while it is n for the other measures under reasonable assumptions on the class of loss functions. We begin with the study of two particular cases to illustrate our results.
△ Less
Submitted 5 October, 2004;
originally announced October 2004.
-
Scarred Patterns in Surface Waves
Authors:
A. Kudrolli,
Mathew C. Abraham,
J. P. Gollub
Abstract:
Surface wave patterns are investigated experimentally in a system geometry that has become a paradigm of quantum chaos: the stadium billiard. Linear waves in bounded geometries for which classical ray trajectories are chaotic are known to give rise to scarred patterns. Here, we utilize parametrically forced surface waves (Faraday waves), which become progressively nonlinear beyond the wave insta…
▽ More
Surface wave patterns are investigated experimentally in a system geometry that has become a paradigm of quantum chaos: the stadium billiard. Linear waves in bounded geometries for which classical ray trajectories are chaotic are known to give rise to scarred patterns. Here, we utilize parametrically forced surface waves (Faraday waves), which become progressively nonlinear beyond the wave instability threshold, to investigate the subtle interplay between boundaries and nonlinearity. Only a subset (three main types) of the computed linear modes of the stadium are observed in a systematic scan. These correspond to modes in which the wave amplitudes are strongly enhanced along paths corresponding to certain periodic ray orbits. Many other modes are found to be suppressed, in general agreement with a prediction by Agam and Altshuler based on boundary dissipation and the Lyapunov exponent of the associated orbit. Spatially asymmetric or disordered (but time-independent) patterns are also found even near onset. As the driving acceleration is increased, the time-independent scarred patterns persist, but in some cases transitions between modes are noted. The onset of spatiotemporal chaos at higher forcing amplitude often involves a nonperiodic oscillation between spatially ordered and disordered states. We characterize this phenomenon using the concept of pattern entropy. The rate of change of the patterns is found to be reduced as the state passes temporarily near the ordered configurations of lower entropy. We also report complex but highly symmetric (time-independent) patterns far above onset in the regime that is normally chaotic.
△ Less
Submitted 14 September, 2000; v1 submitted 24 February, 2000;
originally announced February 2000.
