Zum Inhalt springen
Technische Universität München

Master-Seminar - Deep Learning in Physics (IN2107, IN0014)

Prof. Dr. Nils Thuerey , Nilam Tathawadekar , Shuvayan Brahmachary

Master-Seminar - Deep Learning in Physics (IN2107, IN0014)

Time, PlaceWednesdays 14:00-16:00 in room: MI 02.13.010
Begin19th., October 2022
PrerequisitesIntroduction to Deep Learning

Information

https://ge.in.tum.de/teaching/seminars-ws-20-21 

Content

Using deep learning methods for physical problems is a very quickly developing area of research. The research group of Prof. Thuerey has studied learning-based methods for Navier-Stokes problems and fluid flow applications in recent years, examples of which include learning latent-spaces for physical predictions, generative adversarial networks with temporal coherence, and the inference of Reynolds-averaged Navier-Stokes flows around airfoils. Beyond these physics-based deep learning works of the Thuerey group, this seminar will give an overview of recent developments in the field.

In this course, students will autonomously investigate recent research about machine learning techniques in the physical simulation area. Independent investigation for further reading, critical analysis, and evaluation of the topic are required.

Requirements

Participants are required to first read the assigned paper and start writing a report. This will help you prepare for your presentation.
Attendance
  • It is only allowed to miss a single time-slot. Missing a second one means failing the seminar. If you have to miss any, please let us know in advance.
Advisor
  • An advisor is assigned to each one with the paper.
  • Two weeks before the talk there will be a mandatory meeting with your advisor to review the report and discuss the structure of the presentation.
Report
  • A short report (4 pages max., excluding references in the ACM SIGGRAPH TOG format (acmtog) - you can download the precompiled latex template) should be prepared before the meeting with the advisor.
  • Guideline: You can begin with writing a summary of the work you present as a start point; but, it would be better if you focus more on your own research rather than just finishing with the summary of the paper. We, including you, are not interested in revisiting the work done before; it is more meaningful if you make an effort to put your own reasoning about the work, such as pros and cons, limitation, possible future work, your own ideas for the issues, etc.
Presentation (slides)
  • The participants have to present their topics in a talk (in English), which should last 30 minutes. Don't put too many technical details into the talk, make sure the audience gets the paper's main idea. Be prepared to answer questions regarding the technical details, you could prepare backup slides for that.
  • Afterwards, a short discussion session will follow.
  • Plagiarism is important; please do not simply copy the original authors' slides. You can certainly refer to them.
  • The semi-final slides (PDF) should be sent one week before the talk; otherwise, the talk will be canceled. We strongly encourage you to finalize the semi-final version as far as possible. We will take a look at the version and give feedback. You can revise your slides until your presentation.
  • Be ready in advance. We suggest testing the machines you are going to use before the lecture starts. You can bring your laptop or ask us one (also any converter you need for the projector) in advance. A laser pointer will be provided, so you can use if you want.
  • The final slides and report should be sent after the talk.

Preliminary Schedule

Paper list sent via Email

Send preferred 5 topics/papers by 05 September 2021 (Sunday 23:59)

- 20.10 Kickoff (Introduction lecture)

- 27.10 No meeting

- 3.11 No meeting

- 10.11 Meeting #1 (First presentation) 

- 17.11 Meeting #2

- 24.11 Meeting #3

- 1.12 Meeting #4

- 8.12 Meeting #5

- 15.12 Meeting #6

- 22.12 No meeting

- 12.1 Meeting #7

- 19.1 Meeting #8

- 26.1 Meeting #9 (if necessary!)

Paper list

No.PaperDateFirst nameLast nameSupervisor
1Combining Differentiable PDE Solvers and Graph Neural Networks for Fluid Flow Prediction10 November LouisDorge de Vacher de Saint GéranChen
2Machine learning accelerated computational fluid dynamics10 November TianhaoLinHolzschuh
3Assessment of unsteady flow predictions using hybrid deep learning based reduced-order models    
4Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solvers17 NovemberAnnalenaKoflerChen
5Deep learning methods for super-resolution reconstruction of turbulent flows    
6Transfer learning for nonlinear dynamics and its application to fluid turbulence https://arxiv.org/pdf/2009.01407.pdf17 NovemberEgeAhmedHolzschuh
7Learning to control PDEs with differentiable physics https://arxiv.org/pdf/2001.07457.pdf24 NovemberAnisYaichChen
8Discovering physical concepts with neural networks https://arxiv.org/pdf/1807.10300.pdf24 NovemberAjlaKarisikHolzschuh
9Neural Ordinary Differential Equations https://arxiv.org/pdf/1806.07366.pdf1 DecemberRobertHajdaChen
10Hamiltonian Neural Networks1 DecemberJosé MiguelFerreira HenriquesHolzschuh
11Physics Informed Deep Learning: Data-driven Solutions of Nonlinear Partial Differential Equations  https://arxiv.org/pdf/1711.10561.pdf8 DecemberChristinaNuss-BrillHolzschuh
12Model identification of reduced order fluid dynamics systems using deep learning    
13Deep learning for universal linear embeddings of nonlinear dynamics https://www.nature.com/articles/s41467-018-07210-0.pdf8 DecemberQingSunChen
14tempoGAN: A Temporally Coherent, Volumetric GAN for Super-resolution Fluid Flow https://arxiv.org/pdf/1801.0971015 DecemberMertÜlkerChen
15Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations15 DecemberStefanFrischHolzschuh
16Lagrangian Fluid Simulation with Continuous Convolutions https://openreview.net/pdf?id=B1lDoJSYDH    
17Solving high-dimensional partial differential equations using deep learning https://www.pnas.org/content/115/34/8505    
18Learning data-driven discretizations for partial differential equations https://www.pnas.org/content/pnas/116/31/15344.full.pdf    
19SPNets: Differentiable Fluid Dynamics for Deep Neural Networks https://arxiv.org/pdf/1806.06094.pdf12 Jan. 2022Anirudh NarayananBalaramanChen
20Augmenting Physical Models with Deep Networks for Complex Dynamics Forecasting https://arxiv.org/abs/2010.0445612 Jan. 2022TobiasNeumeierHolzschuh
21Discovering Symbolic Models from Deep Learning with Inductive Biases https://arxiv.org/pdf/2006.11287.pdf19 Jan. 2022AndreasStöckelerChen
22Differentiable Strong Lensing: Uniting Gravity and Neural Nets through Differentiable Probabilistic Programming https://arxiv.org/pdf/1910.06157.pdf19 Jan. 2022NadiiaMatskoHolzschuh
23DeepONet: Learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operators arxiv.org/abs/1910.03193    

Papers

DatePresenterPaper
Nov. 20Sagar Garg2016, Tompson et al., Accelerating Eulerian Fluid Simulation With Convolutional Networks, arXiv.org
Dec. 11Konrad Eder2018, Schenck, SPNets: Differentiable Fluid Dynamics for Deep Neural Networks, arXiv.org
--2015, Ladicky et al., Data-driven Fluid Simulations using Regression Forests, ACM Trans. Graph.
Nov. 27Stephen Ryan2018, Kim et al., Deep Fluids: A Generative Network for Parameterized Fluid Simulations, arXiv.org
--2016, Holden et al., A Deep Learning Framework for Character Motion Synthesis and Editing, ACM Trans. Graph.
Nov. 6Achraf Aroua2018, Pathak et al., Model-Free Prediction of Large Spatiotemporally Chaotic Systems from Data: A Reservoir Computing Approach, Physical review letters
Dec. 18Keerthi Gaddameedi2018, Y. Xie et al., tempoGAN: A Temporally Coherent, Volumetric GAN for Super-resolution Fluid Flow, ACM Trans. Graph.
Nov. 20Mohamed Attia2016, C. Yang et al., Data-driven projection method in fluid simulation, Computer Animation and Virtual Worlds
Nov. 27Jan Luca Watter2018, Bailey et al., Fast and deep deformation approximations, ACM Trans. Graph.
Dec. 11Amir Nourinia2019, Li et al., Learning Particle Dynamics for Manipulating Rigid Bodies, Deformable Objects, and Fluids, ICLR 2019 Conference
Nov. 13Max Oberberger2019, Thuerey et al., Deep Learning Methods for Reynolds-Averaged Navier-Stokes Simulations of Airfoil Flows, arXiv.org
Dec. 18Shucheng Yang2017, Um et al., Liquid Splash Modeling with Neural Networks, Computer Graphics Forum
Nov. 6Anna Maria Geissinger2019, Raissi et al. Physics-informed Neural Networks: A Deep Learning Framework for Solving Forward and Inverse Problems Involving Nonlinear Partial Differential Equations, Journal of Computational Physics
--2019, Zhu et al., Machine learning Methods for Turbulence Modeling in Subsonic Flows Around Airfoils, AIP
Nov. 13Roland Konlechner2018, Zhang et al., Application of Convolutional Neural Network to Predict Airfoil Lift Coefficient, arXiv.org
--2019, Raissi et al., Deep learning of vortex-induced vibrations, Journal of Fluid Mechanics
Dec. 4David Elias Drothier2019, Hu et al., ChainQueen: A Real-Time Differentiable Physical Simulator for Soft Robotics, arXiv.org
Dec. 4Anjie Guo2019, Xu et al., DensePhysNet: Learning Dense Physical Object Representations via Multi-step Dynamic Interaction, RSS 2019

You can access the papers through TUM library's eAccess.

Papers

Discovering Nonlinear PDEs from Scarce Data with Physics-encoded Learning

Model identification of reduced order fluid dynamics systems using deep learning

Direct shape optimization through deep reinforcement learning

Numerical investigation of minimum drag profiles in laminar flow using deep learning surrogates

Data-driven nonlinear aeroelastic models of morphing wings for control

Accelerating Eulerian Fluid Simulation With Convolutional Networks

Deep Learning Methods for Reynolds-Averaged Navier-Stokes Simulations of Airfoil Flows

Deep learning methods for super-resolution reconstruction of turbulent flows

Physics Informed Deep Learning: Data-driven Solutions of Nonlinear Partial Differential Equations

Learning data-driven discretizations for partial differential equations

Neural Ordinary Differential Equations

Learning to control PDEs with differentiable physics

Combining Differentiable PDE Solvers and Graph Neural Networks for Fluid Flow Prediction

CFDNet: a deep learning-based accelerator for fluid simulations

Machine learning accelerated computational fluid dynamics

Learned Coarse Models for Efficient Turbulence Simulation

Lagrangian Fluid Simulation with Continuous Convolutions

Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solvers

Solving high-dimensional partial differential equations using deep learning

Transfer learning for nonlinear dynamics and its application to fluid turbulence

Towards Physics-informed Deep Learning for Turbulent Flow Prediction

Message Passing Neural PDE Solvers

Neural Solvers for Fast and Accurate Numerical Optimal Control

Presentations

Presentation datePresenterPaperAdvisor
09.11.2022K. BaliAccelerating Eulerian Fluid Simulation With Convolutional NetworksNilam
09.11.2022K. SelimData-driven nonlinear aeroelastic models of morphing wings for controlNilam
16.11.2022A. PimpalkarLagrangian Fluid Simulation with Continuous ConvolutionsShuvayan
16.11.2022Y. ShehataDirect shape optimization through deep reinforcement learningShuvayan
23.11.2022J. OldenstädtPhysics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential EquationsShuvayan
23.11.2022G. GruhlkeLearning data-driven discretizations for partial differential equationsNilam
30.11.2022S. JavadDeep Learning Methods for Reynolds-Averaged Navier-Stokes Simulations Shuvayan
30.11.2022D. GrotheDeep learning methods for super-resolution reconstruction of turbulent flowsNilam
07.12.2022X. LuoNeural Ordinary Differential EquationsNilam
07.12.2022A. ChumakLearning to control PDEs with differentiable physicsShuvayan
14.12.2022M. Saeidy PourMachine learning accelerated computational fluid dynamicsShuvayan
14.12.2022G. JingCFDNet: a deep learning-based accelerator for fluid simulations Shuvayan
21.12.2022J. LassenTowards physics-informed deep learning for turbulent flow predictionShuvayan
21.12.2022G. KripkaCombining Differentiable PDE Solvers and Graph Neural Networks for Fluid Flow Prediction Nilam
11.01.2023F. BrandisLearned Coarse Models for Efficient Turbulence SimulationNilam
11.01.2023S. MohamedSolving high-dimensional partial differential equations using deep learningNilam

References

To top