-
FaçAID: A Transformer Model for Neuro-Symbolic Facade Reconstruction
Authors:
Aleksander Plocharski,
Jan Swidzinski,
Joanna Porter-Sobieraj,
Przemyslaw Musialski
Abstract:
We introduce a neuro-symbolic transformer-based model that converts flat, segmented facade structures into procedural definitions using a custom-designed split grammar. To facilitate this, we first develop a semi-complex split grammar tailored for architectural facades and then generate a dataset comprising of facades alongside their corresponding procedural representations. This dataset is used t…
▽ More
We introduce a neuro-symbolic transformer-based model that converts flat, segmented facade structures into procedural definitions using a custom-designed split grammar. To facilitate this, we first develop a semi-complex split grammar tailored for architectural facades and then generate a dataset comprising of facades alongside their corresponding procedural representations. This dataset is used to train our transformer model to convert segmented, flat facades into the procedural language of our grammar. During inference, the model applies this learned transformation to new facade segmentations, providing a procedural representation that users can adjust to generate varied facade designs. This method not only automates the conversion of static facade images into dynamic, editable procedural formats but also enhances the design flexibility, allowing for easy modifications.
△ Less
Submitted 13 September, 2024; v1 submitted 3 June, 2024;
originally announced June 2024.
-
Relativistic hydrodynamics on graphics processing units
Authors:
Jan Sikorski,
Sebastian Cygert,
Joanna Porter-Sobieraj,
Marcin Słodkowski,
Piotr Krzyżanowski,
Natalia Książek,
Przemysław Duda
Abstract:
Hydrodynamics calculations have been successfully used in studies of the bulk properties of the Quark-Gluon Plasma, particularly of elliptic flow and shear viscosity. However, there are areas (for instance event-by-event simulations for flow fluctuations and higher-order flow harmonics studies) where further advancement is hampered by lack of efficient and precise 3+1D~program. This problem can be…
▽ More
Hydrodynamics calculations have been successfully used in studies of the bulk properties of the Quark-Gluon Plasma, particularly of elliptic flow and shear viscosity. However, there are areas (for instance event-by-event simulations for flow fluctuations and higher-order flow harmonics studies) where further advancement is hampered by lack of efficient and precise 3+1D~program. This problem can be solved by using Graphics Processing Unit (GPU) computing, which offers unprecedented increase of the computing power compared to standard CPU simulations. In this work, we present an implementation of 3+1D ideal hydrodynamics simulations on the Graphics Processing Unit using Nvidia CUDA framework. MUSTA-FORCE (MUlti STAge, First ORder CEntral, with a~slope limiter and MUSCL reconstruction) and WENO (Weighted Essentially Non-Oscillating) schemes are employed in the simulations, delivering second (MUSTA-FORCE), fifth and seventh (WENO) order of accuracy. Third order Runge-Kutta scheme was used for integration in the time domain. Our implementation improves the performance by about 2~orders of magnitude compared to a~single threaded program. The algorithm tests of 1+1D~shock tube and 3+1D~simulations with ellipsoidal and Hubble-like expansion are presented.
△ Less
Submitted 12 April, 2016;
originally announced April 2016.
-
Jet-induced modifications of the characteristic of the bulk nuclear matter
Authors:
P. Marcinkowski,
M. Słodkowski,
D. Kikoła,
J. Sikorski,
J. Porter-Sobieraj,
P. Gawryszewski,
B. Zygmunt
Abstract:
We present our studies on jet-induced modifications of the characteristic of the bulk nuclear matter. To describe such a matter, we use efficient relativistic hydrodynamic simulations in (3+1) dimensions employing the Graphics Processing Unit (GPU) in the parallel programming framework. We use Cartesian coordinates in the calculations to ensure a high spatial resolution that is constant throughout…
▽ More
We present our studies on jet-induced modifications of the characteristic of the bulk nuclear matter. To describe such a matter, we use efficient relativistic hydrodynamic simulations in (3+1) dimensions employing the Graphics Processing Unit (GPU) in the parallel programming framework. We use Cartesian coordinates in the calculations to ensure a high spatial resolution that is constant throughout the evolution of the system. We show our results on how jets modify the hydrodynamics fields and discuss the implications.
△ Less
Submitted 2 November, 2015;
originally announced November 2015.
-
A MUSTA-FORCE algorithm for solving partial differential equations of relativistic hydrodynamics
Authors:
J. Porter-Sobieraj,
M. Słodkowski,
D. Kikoła,
J. Sikorski,
P. Aszklar
Abstract:
Understanding event-by-event correlations and fluctuations is crucial for the comprehension of the dynamics of heavy ion collisions. Relativistic hydrodynamics is an elegant tool for modeling these phenomena; however, such simulations are time-consuming, and conventional CPU calculations are not suitable for event-by-event calculations. This work presents a feasibility study of a new hydrodynamic…
▽ More
Understanding event-by-event correlations and fluctuations is crucial for the comprehension of the dynamics of heavy ion collisions. Relativistic hydrodynamics is an elegant tool for modeling these phenomena; however, such simulations are time-consuming, and conventional CPU calculations are not suitable for event-by-event calculations. This work presents a feasibility study of a new hydrodynamic code that employs graphics processing units together with a general MUSTA-FORCE algorithm (Multi-Stage Riemann Algorithm - First Order Centered scheme) to deliver a high-performance yet universal tool for event-by-event hydrodynamic simulations. We also investigate the performance of selected slope limiters that reduce the amount of numeric oscillations and diffusion in the presence of strong discontinuities and shock waves. The numerical results are compared to the exact solutions to assess the code's accuracy.
△ Less
Submitted 16 February, 2018; v1 submitted 21 October, 2015;
originally announced October 2015.
