Christos Lataniotis


Last Name

Lataniotis

First Name

Christos

ORCID

0000-0002-2888-4050

Organisational unit

Filters

2015 - 2019162020 - 20236
Reset filters

Search Results

Publications 1 - 10 of 22
  • Uncertainty Quantification in the cloud with UQCloud
    Item type: Conference Paper
    Lataniotis, Christos; Marelli, Stefano; Sudret, Bruno (2021)
    General-purpose uncertainty quantification software has become a well established requirement in modern engineering workflows. Different communities (e.g. applied maths, engineering, economics, etc.), however, generally employ diverse arrays of technologies and workflows, from computing infrastructure to programming languages. To overcome the intrinsic limitation of a single language, standalone software package, we introduce UQCloud, an OS- and programming language- independent, cloud-based version of UQLab. UQCloud follows a software-asa-service (SaaS) model, that allows anyone to take advantage of the well-established UQLab suite, without the need to adapt their computational workflows to include Matlab.
  • An adaptive algorithm based on spectral likelihood expansion for efficient Bayesian calibration
    Item type: Other Conference Item
    Wagner, Paul-Remo; Lataniotis, Christos; Marelli, Stefano; et al. (2019)
  • Combining machine learning and surrogate modeling for data-driven uncertainty propagation in high-dimension
    Item type: Other Conference Item
    Lataniotis, Christos; Marelli, Stefano; Sudret, Bruno (2019)
  • Extending classical surrogate modelling to ultrahigh dimensional problems through supervised dimensionality reduction: a data-driven approach
    Item type: Working Paper
    Lataniotis, Christos; Marelli, Stefano; Sudret, Bruno (2018)
    arXiv
    Thanks to their versatility, ease of deployment and high-performance, surrogate models have become staple tools in the arsenal of uncertainty quantification (UQ). From local interpolants to global spectral decompositions, surrogates are characterised by their ability to efficiently emulate complex computational models based on a small set of model runs used for training. An inherent limitation of many surrogate models is their susceptibility to the curse of dimensionality, which traditionally limits their applicability to a maximum of O(10^2) input dimensions. We present a novel approach at high-dimensional surrogate modelling that is model-, dimensionality reduction- and surrogate model- agnostic (black box), and can enable the solution of high dimensional (i.e. up to O(10^4)) problems. After introducing the general algorithm, we demonstrate its performance by combining Kriging and polynomial chaos expansions surrogates and kernel principal component analysis. In particular, we compare the generalisation performance that the resulting surrogates achieve to the classical sequential application of dimensionality reduction followed by surrogate modelling on several benchmark applications, comprising an analytical function and two engineering applications of increasing dimensionality and complexity.
  • UQLab & UQ[py]Lab - project updates and outlook
    Item type: Other Conference Item
    Hlobilová, Adéla; Lataniotis, Christos; Marelli, Stefano; et al. (2023)
    The adoption of advanced uncertainty quantification techniques is steadily increasing throughout the academic and industrial applications landscape. The assessment of uncertainty in technological systems of societal relevance (e.g. in civil and aerospace engineering) is being gradually mandated, or at least included in engineering construction codes all around the world. In parallel, the design and maintenance of renewable energy systems (e.g. wind turbine generators) and related policies crucially depend on the highly volatile socio-economic and climate-related conditions. As a result, the need for powerful and accessible uncertainty quantification software tools keep growing. For almost ten years, the UQLab project has been dedicated to the development and dissemination of an open source software platform for uncertainty quantification in engineering and applied sciences. Counting now close to 6,000 users, UQLab has achieved widespread recognition as one of the reference software in UQ, despite being based on Matlab, a relatively less popular platform. To further boost the dissemination potential of such a tool, in 2020 we launched the UQCloud project, which provides a platform- and language- independent cloud version of UQLab, followed by the first beta release of its python client, UQ[py]Lab (https://uqpylab.uq-cloud.io). In this contribution, we will officially present the first public release of UQ[py]Lab 1.0, which provides a fairly complete UQLab experience in python, and paves the way to future clients in additional languages, such as R and Julia. We will also provide an overview of the intertwined development roadmap of both UQLab and UQ[py]Lab, with an eye on the novel scenarios made possible by the unique features of a software-as-a-service (SaaS) platform.
  • Sparse polynomial chaos expansions as a machine learning regression technique
    Item type: Other Conference Item
    Sudret, Bruno; Marelli, Stefano; Lataniotis, Christos (2015)
  • Combining feature mapping and Gaussian process modelling in the context of UQ
    Item type: Other Conference Item
    Lataniotis, Christos; Marelli, Stefano; Sudret, Bruno (2016)
  • Dimensionality reduction and surrogate modelling for high-dimensional UQ problems
    Item type: Other Conference Item
    Lataniotis, Christos; Marelli, Stefano; Sudret, Bruno (2018)
  • Stochastic spectral embedding
    Item type: Journal Article
    Marelli, Stefano; Wagner, Paul-Remo; Lataniotis, Christos; et al. (2021)
    International Journal for Uncertainty Quantification
    Constructing approximations that can accurately mimic the behavior of complex models at reduced computational costs is an important aspect of uncertainty quantification. Despite their flexibility and efficiency, classical surrogate models such as kriging or polynomial chaos expansions tend to struggle with highly nonlinear, localized, or nonstationary computational models. We hereby propose a novel sequential adaptive surrogate modeling method based on recursively embedding locally spectral expansions. It is achieved by means of disjoint recursive partitioning of the input domain, which consists in sequentially splitting the latter into smaller subdomains, and constructing simpler local spectral expansions in each, exploiting the trade-off complexity vs. locality. The resulting expansion, which we refer to as "stochastic spectral embedding" (SSE), is a piecewise continuous approximation of the model response that shows promising approximation capabilities, and good scaling with both the problem dimension and the size of the training set. We finally show how the method compares favorably against state-of-the-art sparse polynomial chaos expansions on a set of models with different complexity and input dimension. © 2021 by Begell House, Inc.
  • Dimensionality reduction and surrogate modelling for high-dimensional UQ problems
    Item type: Other Conference Item
    Lataniotis, Christos; Marelli, Stefano; Sudret, Bruno (2017)
Publications 1 - 10 of 22