Name: 7th Round Table on Deep Learning @DESY
Start: 2024-11-22T09:10:00+01:00
End: 2024-11-22T16:20:00+01:00
Location: DESY

7th Round Table on Deep Learning @DESY

Friday 22 November 2024 - 09:10

Monday 18 November 2024
Tuesday 19 November 2024
Wednesday 20 November 2024
Thursday 21 November 2024
Friday 22 November 2024
09:30
09:30 - 09:50
Room: Flash Seminar Room
Contributions

09:30 Welcome - Philipp Heuser (DESY/Helmholtz Imaging)

09:35 The Helmholtz Model Zoo - Philipp Heuser (DESY/Helmholtz Imaging)

09:40 LEAPS special interest group on AI/ML - Alexander Paul Millar (IT (Research and Innovation in Scientific Co))

09:45 Hardware for AI at DESY - Yves Kemp (IT (IT Systems))
09:50 Deep Learning in Pelagic Imaging: Advancements and GEOMAR’s Role in the HFMI AqQua Project - Martin Schröder (GEOMAR)
Deep Learning in Pelagic Imaging: Advancements and GEOMAR’s Role in the HFMI AqQua Project
- Martin Schröder (GEOMAR)
09:50 - 10:30
Room: Flash Seminar Room Pelagic imaging, the capture of images of plankton and particles in the open-water zones of the oceans, is central for understanding plankton diversity, distribution, and dynamics on a large scale.The integration of deep learning (DL) into pelagic imaging workflows offers the potential to improve the precision and scalability of image-based analyses in plankton research. This talk will first review our recent work in advancing automated image analysis methods tailored to address the challenges of species classification, trait extraction, and image annotation in plankton studies, focusing on the development of MorphoCluster, a cluster-based, interactive image classification tool. Key results indicate that MorphoCluster surpasses traditional methods in sorting accuracy and annotation speed. Building on this foundation, the talk will outline GEOMAR’s part in the upcoming HFMI AqQua project, a joint initiative of Helmholtz institutes to create a versatile deep learning model for plankton image recognition, trained on billions of images from diverse instruments.
10:30 LLMs and Foundation Model
LLMs and Foundation Model
10:30 - 10:54
Room: Flash Seminar Room
Contributions

10:30 OmniJet-$\alpha$: The first cross-task foundation model for particle physics - Joschka Valentin Maria Birk (None)

10:36 GenAI@IT - Tim Moeller (IT (IT Information Fabrics))

10:42 Development of a Professional Q&A AI Agent for Beam Scientists and Users Using LLM and RAG Technologies - Zhongzheng Zhou (FS-EC (Experimente Control))

10:48 Enhancing Accelerator Operation with Large Language Models - Frank Mayet (MXL (XFEL))
10:54 Coffee
Coffee
10:54 - 11:20
Room: Flash Seminar Room
11:20
11:20 - 12:30
Room: Flash Seminar Room
Contributions

11:20 Collaboration on phase contrast imaging and tomography at PETRA III - Johannes Hagemann (FS-PETRA (PETRA III))

11:26 A Good Phase: Solving the Near-field Phase Retrieval Inverse Problem Quantitatively Using ForwardNET - Dawit Hailu (Hereon (Helmholtz-Zentrum Hereon))

11:32 Quality Estimation of retrieved phases - Engin Eren (DESY / Helmholtz Imaging)

11:38 Near-field holography image Denoising using Dilated Convolutional Neural Networks - Vahid Rahmani (FS-PETRA (PETRA III))

11:44 Deep-Learning based Image Denoising of MPES data - M Zain Sohail (DESY/RWTH Aachen) Dmytro Kutnyakhov (DESY Photon Science)

11:50 Deep learning for the generation of artificially stained 3D virtual histology of bone implants - Sally Irvine (Helmholtz-Zentrum Hereon)

11:56 VolRAFTv2: Volumetric Optical Flow Network for Digital Volume Correlation - Tak Ming Wong (Hereon (Helmholtz-Zentrum Hereon))

12:02 Deep Neural Networks for XFEL pulse monitoring through emission spectra - Alfredo Bellisario (Uppsala University)

12:08 Coulomb explosion as a high-dimensional probe of single molecules - Benoît Richard (CUI, CFEL and UHH)

12:14 The Virtual Spectrometer at EuXFEL - Danilo Enoque Ferreira de Lima (Eur.XFEL (European XFEL))

12:20 Advanced controls and Machine Learning at FLASHForward - Lewis Anthony Boulton (FTX (FTX Fachgruppe AST))
12:30 Lunch
Lunch
12:30 - 13:30
Room: Flash Seminar Room
13:30 Uncertainty Quantification for Neural Networks: Make your model predictions trustworthy - Steve Schmerler (HZDR)
Uncertainty Quantification for Neural Networks: Make your model predictions trustworthy
- Steve Schmerler (HZDR)
13:30 - 14:30
Room: Flash Seminar Room In machine learning, the ability to make reliable predictions is paramount. Yet, standard ML models and pipelines provide only point predictions without accounting for model confidence (or the lack thereof). Uncertainty in model outputs, especially when faced with out-of-distribution (OOD) data, is essential when deploying models in production. This talk serves as an introduction to the concepts and techniques for quantifying uncertainty in machine learning models. We will explore the different sources of uncertainty and cover various methods for estimating these uncertainties effectively. By understanding and addressing uncertainty, particularly in the context of OOD data, practitioners can enhance the robustness of their models and foster greater confidence in model predictions.
14:30 Coffee
Coffee
14:30 - 14:50
Room: Flash Seminar Room
14:50
14:50 - 15:50
Room: Flash Seminar Room
Contributions

14:50 Optimising operation of European XFEL with artificial intelligence - David Doblas Jimenez (Eur.XFEL (European XFEL))

14:56 A Surrogate Model for Data-Driven Prediction of X-Ray Pulse Properties at the European XFEL - Fady Adibsamy Bishara (Eur.XFEL (European XFEL))

15:02 Optimizing Free-Electron Laser Operation with Neural Network-Based Beam Pointing Predictions - Farzad Jafarinia (MXL (XFEL))

15:08 Real-time ML-based anomaly detection with FPGAs at the LHC - Artur Lobanov (Universität Hamburg) Finn Jonathan Labe (Universität Hamburg)

15:14 Neural Quasiprobabilistic Likelihood Ratio Estimation with Negatively Weighted Data - Stephen Jiggins (ATLAS (ATLAS-Experiment))

15:20 Design choices for Flow Models in HEP data - Henry Day-Hall (FTX (FTX Fachgruppe SFT))

15:26 Anomaly searches in Atlas track distributions - Max Fuste Costa (ATLAS (ATLAS-Experiment))

15:32 CaloClouds III: Ultra-Fast Geometry-Independent Highly-Granular Calorimeter Simulation - Anatolii Korol (FTX (FTX Fachgruppe SFT))

15:38 Higher Resolution and Angular Conditioning for Normalizing-Flow-based Generation of Calorimeter Showers - Thorsten Buss (Universität Hamburg)

15:44 Quantum Machine Learning for Calorimeter Simulation - Kerstin Borras (DESY and RWTH Aachen University)

Choose timezone