-
Prediction and identification of point defect fingerprints in the X-ray photoelectron spectra of TiN$_x$
Authors:
Pavel Ondračka,
Pauline Kümmerl,
Marcus Hans,
Stanislav Mráz,
Daniel Primetzhofer,
David Holec,
Petr Vašina,
Jochen M. Schneider
Abstract:
We investigate the effect of selected N and Ti point defects in $B$1 TiN on N 1s and Ti 2p$_{3/2}$ binding energies (BE) by experiments and ab initio calculations. X-ray photoelectron spectroscopy (XPS) measurements of Ti-deficient TiN films reveal additional N 1s spectral components at lower binding energies. Ab initio calculations predict that these components are caused by either Ti vacancies,…
▽ More
We investigate the effect of selected N and Ti point defects in $B$1 TiN on N 1s and Ti 2p$_{3/2}$ binding energies (BE) by experiments and ab initio calculations. X-ray photoelectron spectroscopy (XPS) measurements of Ti-deficient TiN films reveal additional N 1s spectral components at lower binding energies. Ab initio calculations predict that these components are caused by either Ti vacancies, which induce a N 1s BE shift of $-0.53$ eV in its first N neighbors, and/or N tetrahedral interstitials, which have their N 1s BE shifted by $-1.18$ eV and also shift BE of their first N neighbors by $-0.53$ eV. However, the {\it ab initio} calculations also reveal that the tetrahedral N interstitial is unstable at room temperature. We, therefore, unambiguously attribute the detected signal to Ti vacancies. Furthermore, the vacancy concentration in Ti-deficient TiN was quantified with XPS supported by ab initio calculations. The largest BE shifts of $-1.53$, $-1.80$ and $-2.28$ eV for Ti 2p$_{3/2}$ electrons are predicted for the Ti tetrahedral, split (10$\overline{1}$)-aligned and split (111)-aligned interstitial atoms, respectively, and we, therefore, propose XPS could detect them. Other defects such as N vacancy or N split (10$\overline{1}$)-aligned interstitial introduce smaller N 1s and Ti 2p$_{3/2}$ BE shifts and are unlikely to be detectable experimentally. Our work highlights the potential of ab initio-guided XPS measurements in detecting and quantifying point defects in $B$1 TiN.
△ Less
Submitted 28 March, 2024;
originally announced March 2024.
-
Valence electron concentration- and N vacancy-induced elasticity in cubic early transition metal nitrides
Authors:
Soheil Karimi Aghda,
Dimitri Bogdanovski,
Lukas Loefler,
Heng Han Sua,
Lena Patterer,
Damian M. Holzapfel,
Arnaud le Febvrier,
Marcus Hans,
Daniel Primetzhofer,
Jochen M. Schneider
Abstract:
Motivated by frequently reported deviations from stoichiometry in cubic transition metal nitride (TMNx) thin films, the effect of N-vacancy concentration on the elastic properties of cubic TiNx, ZrNx, VNx, NbNx, and MoNx (0.72<x<1.00) is systematically studied by density functional theory (DFT) calculations. The predictions are validated experimentally for VNx (0.77<x<0.97). The DFT results indica…
▽ More
Motivated by frequently reported deviations from stoichiometry in cubic transition metal nitride (TMNx) thin films, the effect of N-vacancy concentration on the elastic properties of cubic TiNx, ZrNx, VNx, NbNx, and MoNx (0.72<x<1.00) is systematically studied by density functional theory (DFT) calculations. The predictions are validated experimentally for VNx (0.77<x<0.97). The DFT results indicate that the elastic behavior of the TMNx depends on both the N-vacancy concentration and the valence electron concentration (VEC) of the transition metal: While TiNx and ZrNx exhibit vacancy-induced reductions in elastic modulus, VNx and NbNx show an increase. These trends can be rationalized by considering vacancy-induced changes in elastic anisotropy and bonding. While introduction of N-vacancies in TiNx results in a significant reduction of elastic modulus along all directions and a lower average bond strength of Ti-N, the vacancy-induced reduction in [001] direction of VNx is overcompensated by the higher stiffness along [011] and [111] directions, resulting in a higher average bond strength of V-N. To validate the predicted vacancy-induced changes in elasticity experimentally, close-to-single-crystal VNx (0.77<x<0.97) are grown on MgO(001) substrates. As the N-content is reduced, the relaxed lattice parameter a0, as probed by X-ray diffraction, decreases from 4.128 A to 4.096 A. This reduction in lattice parameter is accompanied by an anomalous 11% increase in elastic modulus, as determined by nanoindentation. As the experimental data agree with the predictions, the elasticity enhancement in VNx upon N-vacancy formation can be understood based on the concomitant changes in elastic anisotropy and bonding.
△ Less
Submitted 27 May, 2023;
originally announced May 2023.
-
Bose Einstein condensate as nonlinear block of a Machine Learning pipeline
Authors:
Maurus Hans,
Elinor Kath,
Marius Sparn,
Nikolas Liebster,
Felix Draxler,
Christoph Schnörr,
Helmut Strobel,
Markus K. Oberthaler
Abstract:
Physical systems can be used as an information processing substrate and with that extend traditional computing architectures. For such an application the experimental platform must guarantee pristine control of the initial state, the temporal evolution and readout. All these ingredients are provided by modern experimental realizations of atomic Bose Einstein condensates. By embedding the nonlinear…
▽ More
Physical systems can be used as an information processing substrate and with that extend traditional computing architectures. For such an application the experimental platform must guarantee pristine control of the initial state, the temporal evolution and readout. All these ingredients are provided by modern experimental realizations of atomic Bose Einstein condensates. By embedding the nonlinear evolution of a quantum gas in a Machine Learning pipeline, one can represent nonlinear functions while only linear operations on classical computing of the pipeline are necessary. We demonstrate successful regression and interpolation of a nonlinear function using a quasi one-dimensional cloud of potassium atoms and characterize the performance of our system.
△ Less
Submitted 28 April, 2023;
originally announced April 2023.
-
On the determination of the thermal shock parameter of MAX phases: A combined experimental-computational study
Authors:
Matej Fekete,
Clio Azina,
Pavel Ondračka,
Lukas Löfler,
Dimitri Bogdanovski,
Daniel Primetzhofer,
Marcus Hans,
Jochen M. Schneider
Abstract:
Thermal shock resistance is one of the performance-defining properties for applications where extreme temperature gradients are required. The thermal shock resistance of a material can be described by means of the thermal shock parameter RT. Here, the thermo-mechanical properties required for the calculation of RT are quantum-mechanically predicted, experimentally determined, and compared for Ti3A…
▽ More
Thermal shock resistance is one of the performance-defining properties for applications where extreme temperature gradients are required. The thermal shock resistance of a material can be described by means of the thermal shock parameter RT. Here, the thermo-mechanical properties required for the calculation of RT are quantum-mechanically predicted, experimentally determined, and compared for Ti3AlC2 and Cr2AlC MAX phases. The coatings are synthesized utilizing direct current magnetron sputtering without additional heating, followed by vacuum annealing. It is shown that the RT of both Ti3AlC2 and Cr2AlC obtained via simulations are in good agreement with the experimentally obtained ones. Comparing the MAX phase coatings, both experiments and simulations indicate superior thermal shock behavior of Ti3AlC2 compared to Cr2AlC, attributed primarily to the larger linear coefficient of thermal expansion of Cr2AlC. The results presented herein underline the potential of ab initio calculations for predicting the thermal shock behavior of ionically-covalently bonded materials.
△ Less
Submitted 15 May, 2023; v1 submitted 17 March, 2023;
originally announced March 2023.
-
Chemical and structural characterization of the native oxide scale on a Mg-based alloy
Authors:
Deborah Neuß,
Ingrid E. McCarroll,
Siyuan Zhang,
Eric Woods,
Wassilios J. Delis,
Leandro Tanure,
Hauke Springer,
Stefanie Sandlöbes,
Jing Yang,
Mira Todorova,
Daniela Zander,
Christina Scheu,
Jochen M. Schneider,
Marcus Hans
Abstract:
In this study, the structure and composition of the native oxide forming on the basal plane (0001) of Mg-2Al-0.1Ca is investigated by a correlative approach, combining scanning transmission electron microscopy (STEM) and atom probe tomography (APT). Atom probe specimens were prepared conventionally in a Ga focused ion beam (FIB) as well as a Xe plasma FIB in a cryogenic setup and subsequently clea…
▽ More
In this study, the structure and composition of the native oxide forming on the basal plane (0001) of Mg-2Al-0.1Ca is investigated by a correlative approach, combining scanning transmission electron microscopy (STEM) and atom probe tomography (APT). Atom probe specimens were prepared conventionally in a Ga focused ion beam (FIB) as well as a Xe plasma FIB in a cryogenic setup and subsequently cleaned in the atom probe to remove surface contamination before oxidation. While thermal energy input from the laser and longer atmospheric exposure time increased the measured hydrogen content in the specimen's apex region, cryo preparation revealed, that the hydrogen uptake in magnesium is independent of the employment of conventional or cryogenic FIB preparation. TEM measurements demonstrated the growth of a (111) MgO oxide layer with 3-4 nm thickness on the basal (0001) plane of the Mg atom probe specimen. APT data further revealed the formation of an aluminum-rich region between bulk Mg and the native oxide. The aluminum enrichment of up to ~20 at.% at the interface is consistent with an inward growth of the oxide scale.
△ Less
Submitted 2 June, 2023; v1 submitted 14 March, 2023;
originally announced March 2023.
-
Yttrium incorporation in Cr2AlC: On the metastable phase formation and decomposition of (Cr,Y)2AlC MAX phase thin films
Authors:
Clio Azina,
Tim Bartsch,
Damian M. Holzapfel,
Martin Dahlqvist,
Johanna Rosen,
Lukas Löfler,
Alba San Jose Mendez,
Marcus Hans,
Daniel Primetzhofer,
Jochen M. Schneider
Abstract:
Herein we report on the synthesis of a metastable (Cr,Y)2AlC MAX phase solid solution by co-sputtering from a composite Cr-Al-C and elemental Y target, at room temperature, followed by annealing. While direct high-temperature synthesis resulted in multiphase films, as evidenced by X-ray diffraction analyses, room temperature depositions, followed by annealing to 760 °C led to the formation of phas…
▽ More
Herein we report on the synthesis of a metastable (Cr,Y)2AlC MAX phase solid solution by co-sputtering from a composite Cr-Al-C and elemental Y target, at room temperature, followed by annealing. While direct high-temperature synthesis resulted in multiphase films, as evidenced by X-ray diffraction analyses, room temperature depositions, followed by annealing to 760 °C led to the formation of phase pure (Cr,Y)2AlC by diffusion. Higher annealing temperatures caused decomposition of the metastable phase into Cr2AlC, Y5Al3 , and Cr-carbides. In contrast to pure Cr2AlC, the Y-containing phase crystallizes directly in the MAX phase structure instead of first forming a disordered solid solution. Furthermore, the crystallization temperature was shown to be Y-content dependent and was increased by ~200 °C for 5 at.% Y compared to Cr2AlC. Calculations predicting the metastable phase formation of (Cr,Y)2AlC and its decomposition are in excellent agreement with the experimental findings.
△ Less
Submitted 15 September, 2022;
originally announced September 2022.
-
Curved and expanding spacetime geometries in Bose-Einstein condensates
Authors:
Mireia Tolosa-Simeón,
Álvaro Parra-López,
Natalia Sánchez-Kuntz,
Tobias Haas,
Celia Viermann,
Marius Sparn,
Nikolas Liebster,
Maurus Hans,
Elinor Kath,
Helmut Strobel,
Markus K. Oberthaler,
Stefan Floerchinger
Abstract:
Phonons have the characteristic linear dispersion relation of massless relativistic particles. They arise as low energy excitations of Bose-Einstein condensates and, in nonhomogeneous situations, are governed by a space- and time-dependent acoustic metric. We discuss how this metric can be experimentally designed to realize curved spacetime geometries, in particular, expanding Friedmann-Lemaître-R…
▽ More
Phonons have the characteristic linear dispersion relation of massless relativistic particles. They arise as low energy excitations of Bose-Einstein condensates and, in nonhomogeneous situations, are governed by a space- and time-dependent acoustic metric. We discuss how this metric can be experimentally designed to realize curved spacetime geometries, in particular, expanding Friedmann-Lemaître-Robertson-Walker cosmologies, with negative, vanishing, or positive spatial curvature. A nonvanishing Hubble rate can be obtained through a time-dependent scattering length of the background condensate. For relativistic quantum fields this leads to the phenomenon of particle production, which we describe in detail. We explain how particle production and other interesting features of quantum field theory in curved spacetime can be tested in terms of experimentally accessible correlation functions.
△ Less
Submitted 22 February, 2022; v1 submitted 21 February, 2022;
originally announced February 2022.
-
Quantum field simulator for dynamics in curved spacetime
Authors:
Celia Viermann,
Marius Sparn,
Nikolas Liebster,
Maurus Hans,
Elinor Kath,
Álvaro Parra-López,
Mireia Tolosa-Simeón,
Natalia Sánchez-Kuntz,
Tobias Haas,
Helmut Strobel,
Stefan Floerchinger,
Markus K. Oberthaler
Abstract:
The observed large-scale structure in our Universe is seen as a result of quantum fluctuations amplified by spacetime evolution. This, and related problems in cosmology, asks for an understanding of the quantum fields of the standard model and dark matter in curved spacetime. Even the reduced problem of a scalar quantum field in an explicitly time-dependent spacetime metric is a theoretical challe…
▽ More
The observed large-scale structure in our Universe is seen as a result of quantum fluctuations amplified by spacetime evolution. This, and related problems in cosmology, asks for an understanding of the quantum fields of the standard model and dark matter in curved spacetime. Even the reduced problem of a scalar quantum field in an explicitly time-dependent spacetime metric is a theoretical challenge and thus a quantum field simulator can lead to new insights. Here, we demonstrate such a quantum field simulator in a two-dimensional Bose-Einstein condensate with a configurable trap and adjustable interaction strength to implement this model system. We explicitly show the realisation of spacetimes with positive and negative spatial curvature by wave packet propagation and confirm particle pair production in controlled power-law expansion of space. We find quantitative agreement with new analytical predictions for different curvatures in time and space. This benchmarks and thereby establishes a quantum field simulator of a new class. In the future, straightforward upgrades offer the possibility to enter new, so far unexplored, regimes that give further insight into relativistic quantum field dynamics.
△ Less
Submitted 22 February, 2022; v1 submitted 21 February, 2022;
originally announced February 2022.
-
Microstructure, grain boundary evolution and anisotropic Fe segregation in (0001) textured Ti thin films
Authors:
Vivek Devulapalli,
Marcus Hans,
Prithiv T. Sukumar,
Jochen M. Schneider,
Gerhard Dehm,
Christian H. Liebscher
Abstract:
The structure and chemistry of grain boundaries (GBs) are crucial in determining polycrystalline materials' properties. Faceting and solute segregation to minimize the GB energy is a commonly observed phenomenon. In this paper, a deposition process to obtain pure tilt GBs in titanium (Ti) thin films is presented. By increasing the power density, a transition from polycrystalline film growth to a m…
▽ More
The structure and chemistry of grain boundaries (GBs) are crucial in determining polycrystalline materials' properties. Faceting and solute segregation to minimize the GB energy is a commonly observed phenomenon. In this paper, a deposition process to obtain pure tilt GBs in titanium (Ti) thin films is presented. By increasing the power density, a transition from polycrystalline film growth to a maze bicrystalline Ti film on SrTiO$_3$ (001) substrate is triggered. All the GBs in the bicrystalline thin film are characterized to be $Σ$13 [0001] coincident site lattice (CSL) boundaries. The GB planes are seen to distinctly facet into symmetric {$\bar{7}520$} and asymmetric {$10\bar{1}0$} // {$11\bar{2}0$} segments of 20-50~nm length. Additionally, EDS reveals preferential segregation of iron (Fe) in every alternate symmetric {$\bar{7}520$} segment. Both the faceting and the segregation are explained by a difference in the CSL density between the facet planes. Furthermore, in the GB plane containing Fe segregation, atom probe tomography is used to experimentally determine the GB excess solute to be 1.25~atoms/nm$^{2}$. In summary, the study reveals for the first time a methodology to obtain bicrystalline Ti thin films with strong faceting and anisotropy in iron (Fe) segregation behaviour within the same family of planes.
△ Less
Submitted 8 November, 2021;
originally announced November 2021.
-
$\textit{Ab initio}$-guided X-ray photoelectron spectroscopy quantification of Ti vacancies in Ti$_{1-δ}$O$_x$N$_{1-x}$ thin films
Authors:
Pavel Ondračka,
Marcus Hans,
Damian M. Holzapfel,
Daniel Primetzhofer,
David Holec,
Jochen M. Schneider
Abstract:
$\textit{Ab initio}$ calculations were employed to investigate the effect of oxygen concentration dependent Ti vacancies formation on the core electron binding energy shifts in cubic titanium oxynitride (Ti$_{1-δ}$O$_x$N$_{1-x}$). It was shown, that the presence of a Ti vacancy reduces the 1s core electron binding energy of the first N neighbors by $\sim…
▽ More
$\textit{Ab initio}$ calculations were employed to investigate the effect of oxygen concentration dependent Ti vacancies formation on the core electron binding energy shifts in cubic titanium oxynitride (Ti$_{1-δ}$O$_x$N$_{1-x}$). It was shown, that the presence of a Ti vacancy reduces the 1s core electron binding energy of the first N neighbors by $\sim$0.6 eV and that this effect is additive with respect to the number of vacancies. Hence it is predicted that the Ti vacancy concentration can be revealed from the intensity of the shifted components in the N 1s core spectra region. This notion was critically appraised by fitting the N 1s region obtained via X-ray photoelectron spectroscopy (XPS) measurements of Ti$_{1-δ}$O$_x$N$_{1-x}$ thin films deposited by high power pulsed magnetron sputtering. A model to quantify the Ti vacancy concentration based on the intensity ratio between the N 1s signal components, corresponding to N atoms with locally different Ti vacancy concentration, was developed. Herein a random vacancy distribution was assumed and the influence of surface oxidation from atmospheric exposure after deposition was considered. The so estimated vacancy concentrations are consistent with a model calculating the vacancy concentration based on the O concentrations determined by elastic recoil detection analysis and text book oxidation states and hence electroneutrality. Thus, we have unequivocally established that the formation and population of Ti vacancies in cubic Ti$_{1-δ}$O$_x$N$_{1-x}$ thin films can be quantified by XPS measurements from N 1s core electron binding energy shifts.
△ Less
Submitted 22 October, 2021; v1 submitted 8 October, 2021;
originally announced October 2021.
-
Empirical Bayes Model Averaging with Influential Observations: Tuning Zellner's g Prior for Predictive Robustness
Authors:
Christopher M. Hans,
Mario Peruggia,
Junyan Wang
Abstract:
The behavior of Bayesian model averaging (BMA) for the normal linear regression model in the presence of influential observations that contribute to model misfit is investigated. Remedies to attenuate the potential negative impacts of such observations on inference and prediction are proposed. The methodology is motivated by the view that well-behaved residuals and good predictive performance ofte…
▽ More
The behavior of Bayesian model averaging (BMA) for the normal linear regression model in the presence of influential observations that contribute to model misfit is investigated. Remedies to attenuate the potential negative impacts of such observations on inference and prediction are proposed. The methodology is motivated by the view that well-behaved residuals and good predictive performance often go hand-in-hand. Focus is placed on regression models that use variants on Zellner's g prior. Studying the impact of various forms of model misfit on BMA predictions in simple situations points to prescriptive guidelines for "tuning" Zellner's g prior to obtain optimal predictions. The tuning of the prior distribution is obtained by considering theoretical properties that should be enjoyed by the optimal fits of the various models in the BMA ensemble. The methodology can be thought of as an "empirical Bayes" approach to modeling, as the data help to inform the specification of the prior in an attempt to attenuate the negative impact of influential cases.
△ Less
Submitted 20 November, 2021; v1 submitted 1 March, 2021;
originally announced March 2021.
-
High signal to noise absorption imaging of alkali atoms at moderate magnetic fields
Authors:
Maurus Hans,
Finn Schmutte,
Celia Viermann,
Nikolas Liebster,
Marius Sparn,
Markus K. Oberthaler,
Helmut Strobel
Abstract:
We present an improved scheme for absorption imaging of alkali atoms at moderate magnetic fields, where the excited state is well in the Paschen-Back regime but the ground state hyperfine manifold is not. It utilizes four atomic levels to obtain an approximately closed optical cycle. With the resulting absorption of the corresponding two laser frequencies we extract the atomic column density of a…
▽ More
We present an improved scheme for absorption imaging of alkali atoms at moderate magnetic fields, where the excited state is well in the Paschen-Back regime but the ground state hyperfine manifold is not. It utilizes four atomic levels to obtain an approximately closed optical cycle. With the resulting absorption of the corresponding two laser frequencies we extract the atomic column density of a $^{39}$K Bose-Einstein condensate. The scheme can be readily applied to all other alkali-like species.
△ Less
Submitted 17 February, 2021;
originally announced February 2021.
-
Streaming End-to-End Bilingual ASR Systems with Joint Language Identification
Authors:
Surabhi Punjabi,
Harish Arsikere,
Zeynab Raeesy,
Chander Chandak,
Nikhil Bhave,
Ankish Bansal,
Markus Müller,
Sergio Murillo,
Ariya Rastrow,
Sri Garimella,
Roland Maas,
Mat Hans,
Athanasios Mouchtaris,
Siegfried Kunzmann
Abstract:
Multilingual ASR technology simplifies model training and deployment, but its accuracy is known to depend on the availability of language information at runtime. Since language identity is seldom known beforehand in real-world scenarios, it must be inferred on-the-fly with minimum latency. Furthermore, in voice-activated smart assistant systems, language identity is also required for downstream pr…
▽ More
Multilingual ASR technology simplifies model training and deployment, but its accuracy is known to depend on the availability of language information at runtime. Since language identity is seldom known beforehand in real-world scenarios, it must be inferred on-the-fly with minimum latency. Furthermore, in voice-activated smart assistant systems, language identity is also required for downstream processing of ASR output. In this paper, we introduce streaming, end-to-end, bilingual systems that perform both ASR and language identification (LID) using the recurrent neural network transducer (RNN-T) architecture. On the input side, embeddings from pretrained acoustic-only LID classifiers are used to guide RNN-T training and inference, while on the output side, language targets are jointly modeled with ASR targets. The proposed method is applied to two language pairs: English-Spanish as spoken in the United States, and English-Hindi as spoken in India. Experiments show that for English-Spanish, the bilingual joint ASR-LID architecture matches monolingual ASR and acoustic-only LID accuracies. For the more challenging (owing to within-utterance code switching) case of English-Hindi, English ASR and LID metrics show degradation. Overall, in scenarios where users switch dynamically between languages, the proposed architecture offers a promising simplification over running multiple monolingual ASR models and an LID classifier in parallel.
△ Less
Submitted 8 July, 2020;
originally announced July 2020.
-
Electronic excitation of transition metal nitrides by light ions with keV energies
Authors:
Barbara Bruckner,
Marcus Hans,
Tomas Nyberg,
Grzegorz Greczynski,
Peter Bauer,
Daniel Primetzhofer
Abstract:
We investigated the specific electronic energy deposition by protons and He ions with keV energies in different transition metal nitrides of technological interest. Data were obtained from two different time-of-flight ion scattering setups and show excellent agreement. For protons interacting with light nitrides, i.e. TiN, VN and CrN, very similar stopping cross sections per atom were found, which…
▽ More
We investigated the specific electronic energy deposition by protons and He ions with keV energies in different transition metal nitrides of technological interest. Data were obtained from two different time-of-flight ion scattering setups and show excellent agreement. For protons interacting with light nitrides, i.e. TiN, VN and CrN, very similar stopping cross sections per atom were found, which coincide with literature data of N2 gas for primary energies <= 25 keV. In case of the chemically rather similar nitrides with metal constituents from the 5th and 6th period, i.e. ZrN and HfN, the electronic stopping cross sections were measured to exceed what has been observed for molecular N2 gas. For He ions, electronic energy loss in all nitrides was found to be significantly higher compared to the equivalent data of N2 gas. Additionally, deviations from velocity proportionality of the observed specific electronic energy loss are observed. A comparison with predictions from density functional theory for protons and He ions yields a high apparent efficiency of electronic excitations of the target for the latter projectile. These findings are considered to indicate the contributions of additional mechanisms besides electron hole pair excitations, such as electron capture and loss processes of the projectile or promotion of target electrons in atomic collisions.
△ Less
Submitted 23 January, 2020; v1 submitted 13 September, 2019;
originally announced September 2019.
-
Prediction Using a Bayesian Heteroscedastic Composite Gaussian Process
Authors:
Casey B. Davis,
Christopher M. Hans,
Thomas J. Santner
Abstract:
This research proposes a flexible Bayesian extension of the composite Gaussian process (CGP) model of Ba and Joseph (2012) for predicting (stationary or) non-stationary $y(\mathbf{x})$. The CGP generalizes the regression plus stationary Gaussian process (GP) model by replacing the regression term with a GP. The new model, $Y(\mathbf{x})$, can accommodate large-scale trends estimated by a global GP…
▽ More
This research proposes a flexible Bayesian extension of the composite Gaussian process (CGP) model of Ba and Joseph (2012) for predicting (stationary or) non-stationary $y(\mathbf{x})$. The CGP generalizes the regression plus stationary Gaussian process (GP) model by replacing the regression term with a GP. The new model, $Y(\mathbf{x})$, can accommodate large-scale trends estimated by a global GP, local trends estimated by an independent local GP, and a third process to describe heteroscedastic data in which $Var(Y(\mathbf{x}))$ can depend on the inputs. This paper proposes a prior which ensures that the fitted global mean is smoother than the local deviations, and extends the covariance structure of the CGP to allow for differentially-weighted global and local components. A Markov chain Monte Carlo algorithm is proposed to provide posterior estimates of the parameters, including the values of the heteroscedastic variance at the training and test data locations. The posterior distribution is used to make predictions and to quantify the uncertainty of the predictions using prediction intervals. The method is illustrated using both stationary and non-stationary $y(\mathbf{x})$.
△ Less
Submitted 25 June, 2019;
originally announced June 2019.
-
Comment on Article by Dawid and Musio
Authors:
Christopher M. Hans,
Mario Peruggia
Abstract:
Discussion of "Bayesian Model Selection Based on Proper Scoring Rules" by Dawid and Musio [arXiv:1409.5291].
Discussion of "Bayesian Model Selection Based on Proper Scoring Rules" by Dawid and Musio [arXiv:1409.5291].
△ Less
Submitted 11 May, 2015;
originally announced May 2015.
-
Block Hyper-g Priors in Bayesian Regression
Authors:
Agniva Som,
Christopher M. Hans,
Steven N. MacEachern
Abstract:
The development of prior distributions for Bayesian regression has traditionally been driven by the goal of achieving sensible model selection and parameter estimation. The formalization of properties that characterize good performance has led to the development and popularization of thick tailed mixtures of g priors such as the Zellner--Siow and hyper-g priors. The properties of a particular prio…
▽ More
The development of prior distributions for Bayesian regression has traditionally been driven by the goal of achieving sensible model selection and parameter estimation. The formalization of properties that characterize good performance has led to the development and popularization of thick tailed mixtures of g priors such as the Zellner--Siow and hyper-g priors. The properties of a particular prior are typically illuminated under limits on the likelihood or the prior. In this paper we introduce a new, conditional information asymptotic that is motivated by the common data analysis setting where at least one regression coefficient is much larger than others. We analyze existing mixtures of g priors under this limit and reveal two new behaviors, Essentially Least Squares (ELS) estimation and the Conditional Lindley's Paradox (CLP), and argue that these behaviors are, in general, undesirable. As the driver behind both of these behaviors is the use of a single, latent scale parameter that is common to all coefficients, we propose a block hyper-g prior, defined by first partitioning the covariates into groups and then placing independent hyper-g priors on the corresponding blocks of coefficients. We provide conditions under which ELS and the CLP are avoided by the new class of priors, and provide consistency results under traditional sample size asymptotics.
△ Less
Submitted 13 January, 2015; v1 submitted 24 June, 2014;
originally announced June 2014.
-
Neutral Charm Decays at CLEO: Searches for CP Violation and Mixing
Authors:
R. M. Hans
Abstract:
Recent CLEO results on neutral charm meson decays presented at the XXXVIth Rencontres de Moriond are discussed. We find no evidence of CP asymmetry in five different two-body decay modes of the D0 to pairs of light pseudo-scalar mesons. We present a measurement of the mixing parameter y_CP = -0.011 +- 0.025 +- 0.014 by searching for a lifetime difference between the CP neutral K+ pi- final state…
▽ More
Recent CLEO results on neutral charm meson decays presented at the XXXVIth Rencontres de Moriond are discussed. We find no evidence of CP asymmetry in five different two-body decay modes of the D0 to pairs of light pseudo-scalar mesons. We present a measurement of the mixing parameter y_CP = -0.011 +- 0.025 +- 0.014 by searching for a lifetime difference between the CP neutral K+ pi- final state and the CP even K+ K- and pi+ pi- final states. Finally, we describe the first measurement of the rate of wrong-sign D0 to K+ pi- pi0 decay: R_WS = (0.43 +0.11 -0.10 +- 0.07)%.
△ Less
Submitted 8 May, 2001;
originally announced May 2001.