Machine Learning approaches in Lattice QCD - An interdisciplinary exchange

Feb 27, 2023, 8:00 AM → Mar 3, 2023, 8:00 PM Europe/Berlin

Institute for Advanced Study of the Technische Universität München

Julian Frederic Mayer-Steudte (TUM School of Natural Sciences - T30f), Nora Brambilla (Physik Department, TU Munich), Tom Magorsch (TU Munich), Zeno Kordov (TUM), Antonio Vairo

A workshop for an interdisciplinary exchange for accelerating developments of machine learning techniques in Lattice QCD with experts in Machine learning, Lattice QCD, and other related fields.

---

This is a Covid safe workshop: FFP2 masks will be required.
Please remember to bring with you FFP2 masks

The conference is broadcasted in zoom if you want to join, contact the organizers via email.

Poster.pdf

Monday, February 27

1 Registration

2 Welcome and aims of the workshop

Speaker: Nora Brambilla

3 Machine learning is ubiquitous

Speaker: Maria-Paola Lombardo

Munich_Feb23_MpL.pdf

Generative flow methods

Talks and panel discussion

Convener: Sinead Ryan

4 Aspects of scaling and scalability for flow-based samplers

Speaker: Daniel Hackett

hackett-tum-talk.pdf

5 Learning trivializing flows

Speaker: David Albandea

David_Munich-ML-approaches.pdf

10:45 AM Coffee break

Generative flow methods

Talks and panel discussion

Convener: Sinead Ryan

6 Fourier-Flow model generating Feynman paths

Speaker: Lingxiao Wang

FFM_Munich0227_lingxiao.pdf

7 Stochastic normalizing flows for lattice field theory

Speaker: Elia Cellini

EliaCellini_SNFLFT_ML_approaches.pdf

8 Panel discussion

Speakers: Chair: Sinead Ryan, Members: Simone Bacchio, Daniel Hackett, Maria Paola Lombardo, Alberto Ramos

Panel.pdf

1:00 PM Lunch

ML for particle physics

Talks and panel discussion

Convener: Lukas Heinrich

9 Discovering mathematical structures with neural networks

Speaker: Sven Krippendorf

lattice-workshop.pdf

10 Machine Learning at IceCube

Speaker: Rasmus Orsoe

GNNs in IceCube (Lattice QCD).pdf

11 Improving Normalizing Flows to Sample from Boltzmann Distributions

Speaker: Vincent Stimper

FAB_ML4Lattice_Workshop_Feb_2023.pdf

12 Panel discussion:

Speaker: Chair: Lukas Heinrich

PanelDiscussion.pdf

4:00 PM Coffee break

ML for phase transitions and sign problem mitigation: Contour deformation methods for noise reduction

Talks and panel discussion

Convener: Will Detmold

13 Signal-to-noise improvement with contour deformations

Speaker: Gurtej Kanwar

ContourDeforms_StN_27_02_2023_PDF.pdf

14 Complex normalizing flows and subtractions for sign problems

Speaker: Yukari Yamauchi

YYamauchi.pdf

15 Applying Complex Valued Neural Networks to the Hubbard Model Sign Problem: A Survey and Case Study

Speaker: Marcel Rodekamp

main.pdf

16 Panel discussion: 1. maximum improvement in <sign> 2. ML towards such a maximal improvement 3. Contour deformation for specific observables 4. Translation between observables/theories

1) Is there a way to determine the maximum improvement in <sign> or in S-to-N that is achievable through contour deformations and related
approaches?

2) How can ML be targeted towards such a maximal improvement?

3) Are there features of specific observables that make contour
deformation particularly effective or ineffective?

4) How do contour deformations for one observable/theory translate to other observables/theories?

Speakers: Chair: Will Detmold, Members: Gurtej Kanwar, Andrei Alexandru (The George Washington University), Thomas Luu, Yukari Yamauchi

2023_Munich_ML4Lattice.pdf

Tuesday, February 28

17 Model-Independent Learning of Quantum Phases of Matter with Quantum Convolutional Neural Networks

Speaker: Frank Pollmann

Pollmann_IAS_QCNN.pdf

18 Automatic differentiation for Lattice gauge theories

Speaker: Alberto Ramos

ADML.pdf

10:45 AM Coffee break

Gauge field generation

Talks and panel discussion

Convener: Phiala Shanahan

19 Simulation of the 2D Schwinger Model via machine-learned flows in Global Correction steps

Speaker: Jacob Finkenrath

FlowGC-ML4QCD.pdf

20 Variational Autoregressive Networks for Information Theory

Speaker: Tomasz Stebel

stebel_ML_QCD.pdf

21 Machine-learning-assisted Monte Carlo fails at sampling computationally hard problems

Speaker: Jeanne Trinquier

SAMPLING_PRESENTATION.pdf

22 Panel discussion: 1. Utility of ML for lattice 2. application-specific features of gauge field generation 3. Biggest challenges 4. Applications at an interesting scale

• Temperature check: for each panelist, in 2-5 words, how optimistic or pessimistic do you feel about the utility of ML in accelerating lattice field generation in your context.
• What are the most application-specific features of the LQCD gauge field generation problem i.e., in what important ways is it different from applications in other areas of physics, and how does that impact the way field generation algorithms must be designed?
• What are the biggest challenges to lattice field generation accelerated by ML?
• Circling back to the start: what are your reasons for optimism or pessimism in reaching applications at an interesting scale?

Speakers: Chair: Phiala Shanahan, Members: Fernando Romero Lopez, Jacob Finkenrath, Jeanne Trinquier

04_Panel_2023_Munich_ML4Lattice_configgenPanel.pdf

1:00 PM Group Photo

1:15 PM Lunch

ML for phase transitions and sign problem mitigation

Talks and panel discussion

Convener: Maria-Paola Lombardo

23 Machine learning for quantum field theories with a sign problem

Speaker: Andrei Alexandru

alexandru-ml-latticeqcd-workshop.pdf

24 A machine learning approach to the classification of phase transitions in many flavor QCD

Speaker: Marius Neumann

Neumann_ML4Lattice.pdf

25 Teaching how to teach: defining learning samples to detect phase transitions

Speaker: Francesco Di Renzo

FDR_28feb2023_TUM.pdf

26 Panel discussion: 1. Competitivity of ML for Phase identification 2. ML for Hypothesised Phases 3. Potential of ML for sign porblem

Panel discussion addressing the three main questions:

--Where do we stand with the identification of phases with ML?
Do you feel ML could become competitive with standard techniques for the quantitative studies of known phase transitions i.e. order of the transition at large Nf, chiral limit, etc.

--Which are your views on the need/concrete possibility of identifying hypothesised, not yet detected phase transitions - in particular confinement, axial restoration, anomalous threshold in the QGP

--Can you summarise the status of the sign problem (without ML) (finite baryon density/real time) and how do you assess the potentiality of ML to mitigate it?

Speaker: Chair: Maria Paola Lombardo

ML_PTsign.pdf

4:00 PM Coffee break

Spectral Reconstruction

Convener: Gert Aarts

27 Brief introduction on spectral functions and their computation

Speaker: Gert Aarts

talkmunichML2023.pdf

28 Spectral reconstruction with Gaussian processes

Speaker: Julian Urban

MLQCD-2023-Urban.pdf

29 Spectral reconstruction with neural networks

Speakers: Kai Zhou, Lingxiao Wang

SF_Munich0228_kai_lingxiao.pdf

30 Learning regulators for spectral reconstruction

Speaker: Alexander Rothkopf

Round_Table_SpectralFunc_EXPORT.pdf

31 Panel discussion

Speaker: Chair: Gert Aarts

Wednesday, March 1

32 Introducing Munich Data Science Institute

Speaker: Sylvia Kortüm

EFT, Lattice and ML

Talks and panel discussion

Convener: Nora Brambilla (Physik Department, TU Munich)

33 Learning Trivializing Gradient Flows

Speaker: Simone Bacchio

Bacchio_gflow.pdf

34 An analysis of Bayesian estimates for missing higher orders in perturbative calculations

Speaker: Aleksas Mazeliauskas

mazeliauskas_TUM.pdf

35 Generative models and EFTs

Speaker: Marina Marinkovic

23-03-01_EFTs_for_ML_v3.pdf

10:40 AM Coffee break

EFT, Lattice and ML

Talks and panel discussion

Convener: Nora Brambilla (Physik Department, TU Munich)

36 Finite-volume pionless EFT for multi-nucleon systems with differential programming

Speaker: Fernando Romero-Lopez

munich.pdf

37 Panel discussion: 1. What can EFTs do for ML4LATTICE? 2. What ML4LATTICE could do for EFTs?

--what can EFTs do for ML4LATTICE?
--what ML4LATTICE could do for EFTs?

Speakers: Chair: Nora Brambilla, Members: Marina Marinkovic, Phiala Shanahan, William Detmold, Aleksas Mazeliauskas

panelMLEFT2023.pdf

1:00 PM Lunch

ML as component of exact algorithms

Talks and panel discussion

Convener: Phiala Shanahan

38 Building Transport Maps and Making Good Use of Them

Speaker: Michael Albergo

munich_mar2023.pptx

39 Automatic differentiation for Lattice QCD

Speaker: A. Ramos

ADML2.pdf

40 Symbolic Distillation of Neural Networks

Speaker: Miles Cranmer

interpretable_ml__mar_2023_export.pdf

41 Panel discussion: 1. Meaning of "exact" algorithms 2. in-principle vs in-practice exactness 3. applications for exactness

• What does it mean to have an ML-accelerated algorithm that is “exact”?
Discuss the distinction between exact algs, interpretable algs, and algs that allow error propagation.
• What are the differences between in-principle and in-practice exactness? Are they important?
• In what applications (both in LQFT and drawing parallels to other ares in physics) is it important to guarantee exactness in ML-accelerated algorithms, and where is it unnecessary, impossible, or worth sacrificing?

Speakers: Chair: Phiala Shanahan, Members: Michael Albergo, Andrei Alexandru, Miles Cranmer

2023_Munich_ML4Lattice_exactalgPanel.pdf

4:35 PM Reception

42 How to accelerate gauge field field generation using flow-based and hybrid models

Speaker: Phiala Shanahan

Thursday, March 2

43 ML for particle physics

Speaker: Lukas Heinrich

44 Complex Langevin real-time simulations and ML

Speaker: Alexander Rothkopf

Complex_Langevin_Kernels_EXPORT.pdf

Complex_Langevin_Kernels_EXPORT.pptx

ML for physical interpretation of lattice results

Talks and panel discussion

Convener: Alexander Rothkopf

45 Towards fully bayesian analyses in Lattice QCD

Speaker: Julien Frison

Presentation.pdf

10:50 AM Coffee Break

ML for physical interpretation of lattice results

Talks and panel discussion

Convener: Alexander Rothkopf

46 Physical Concepts from Neural Networks with Two Inputs

Speaker: Sebastian Wetzel

2023MunichWetzel.pdf

47 Ab-Initio Quantum Chemistry via Graph Neural Networks

Speaker: Nicholas Gao

ml4lattice.pptx

48 Panel discussion on ML for physics interpretation

1. Alexander Rothkopf: Setting the stage – ML for physics interpretation
2. Sebastian Wetzel: ML for identifying order parameters & effective d.o.f. in the context of phase transitions
3. Matteo Favoni: ML for identifying defects and effective d.o.f. in the context of confinement
4. Fernando Romero Lopez: ML for extraction of emergent d.o.f. in the context of scattering resonance
5. Miles Cranmer: ML for inference of governing equations in the context of orbital mechanics (remote)
6. Discussion with the audience

Speakers: Chair: Alexander Rothkopf, Members: Sebastian Wetzel, Matteo Favoni, Fernando Romero Lopez, Miles Cranmer

2023MunichInterpretationPanel.pdf

Defects_and_center_vortices.pdf

panel_frl.pdf

Round_Table_ML_PhysicsInterpretation_v2.pdf

1:00 PM Lunch

Symmetry equivariant neural networks/informed neural networks

Talks and panel discussion

Convener: Phiala Shanahan

49 Generation of gauge field configurations with equivariant neural networks

Speaker: Matteo Favoni

Generation_of_gauge_field_configurations_with_equivariant_neural_networks__Munich_.pdf

50 TBA

Speaker: Daniel Hackett

hackett-tum-talk-2.pdf

51 Panel discussion: 1. Symmetry-equivariance vs computational brute force 2. Computational challenges of gauge equivariance

• There is ongoing debate in the ML community about the value of trying to incorporate domain knowledge into architectures vs using computational brute force. What guides your intuition for the value of building symmetry-equivariant architectures? In your experience when is symmetry-equivariance valuable and when is it not?
• What makes gauge equivariance in particular challenging from a computational efficiency perspective?

Speakers: Chair: Phiala Shanahan, Members: Gurtej Kanwar, Christoph Lehner, Matteo Favoni

2023_Munich_ML4Lattice_gaugeequivPanel.pdf

Poster session (With food and drinks)

Friday, March 3

52 ML and quantum field theories

Speaker: Gert Aarts

talkmunichML2023-QFTML.pdf

Inverse Problem

Talks and panel discussion

53 Machine learning hadron spectral functions in Lattice QCD

Speaker: Gabor Papp

München_20230303_1.pptx

54 R-ratio from Lattice QCD

Speaker: Jian Liang

55 Inverse problem solving in nuclear physics with deep learning

Speaker: Kai Zhou

inverse_problem_KaiZhou.pdf

11:00 AM Coffee break

Inverse Problem

Talks and panel discussion

Convener: Andreas Kronfeld

56 On the extraction of hadronic spectral densities from Euclidean correlators

Speaker: Alessandro De Santis

talk.pdf

57 Panel discussion

Speaker: Chair: Andreas Kronfeld

ml4l-inverse-panel.pdf

12:30 PM Lunch

EFT, Lattice and ML

Talks and panel discussion

58 Machine learning for symbolic computation in high energy physics

Speaker: Abdulhakim Alnuqayqdan

59 Alleviating the Sign Problem via Contour Deformation and Machine Learning

Speaker: Thomas Luu

Luu_ML4LQCD_2023.pdf

60 Gauge-equivariant neural networks as preconditioners in lattice QCD

Speaker: Christoph Lehner

talk.pdf

3:15 PM Coffee break

EFT, Lattice and ML

Talks and panel discussion

61 Deep Learning and the Standard Model: a philosophy of science perspective

Speaker: Luigi Scorzato

Deep Learning and the Standard Model - a philosophy of science point of view.pdf

62 The route towards phase transition recognition in Lattice Gauge Theories

Speaker: Andreas Athenodorou

63 Search for Exotic Higgs Bosons using Quantum Machine Learning

Speaker: Nakul Aggarwal

64 Summary

Speaker: Nora Brambilla