Simon Welker
I am a second-year Computer Science PhD Student at University of Hamburg and the Center for Free-Electron Lasers at DESY, supervised by Prof. Timo Gerkmann and Prof. Henry Chapman. My project is part of the Helmholtz graduate school DASHH.
My research focuses on solving inverse problems that arise in the fields of X-ray imaging and speech processing, especially the problem of Phase Retrieval. I aim to find and exploit common ground between both fields of research by viewing these problems through the common lens of signal processing, and to develop novel methods based on both deep learning and classical nonlinear optimization. Currently, I am working on adapting and extending generative diffusion models in order to make solving inverse problems easier with effective generative priors. I am especially interested in the recent topic of Scientific Machine Learning and the confluence of deep learning and differential equations.
Besides my academic studies, I am also a trained software developer and have worked in the industry for 5+ years. I always strive to develop my research code using best practices from my experience in software engineering.
Publications
Analysing Diffusion-based Generative Approaches versus Discriminative Approaches for Speech Restoration
Jean-Marie Lemercier, Julius Richter, Simon Welker, Timo Gerkmann (2023). "Analysing Diffusion-based Generative Approaches versus Discriminative Approaches for Speech Restoration." ICASSP 2023.
DiffPhase: Generative Diffusion-based STFT Phase Retrieval
Tal Peer, Simon Welker, Timo Gerkmann (2023). "DiffPhase: Generative Diffusion-based STFT Phase Retrieval." ICASSP 2023.
StoRM: A Diffusion-based Stochastic Regeneration Model for Speech Enhancement and Dereverberation
Jean-Marie Lemercier, Julius Richter, Simon Welker, Timo Gerkmann (2022). "StoRM: A Diffusion-based Stochastic Regeneration Model for Speech Enhancement and Dereverberation" arXiv preprint arXiv:2212.11851.
Blind Drifting: Diffusion models with a linear SDE drift term for blind image restoration tasks
Simon Welker, Henry N. Chapman, Timo Gerkmann (2022). "Blind Drifting: Diffusion models with a linear SDE drift term for blind image restoration tasks." NeurIPS 2022 Workshop The Symbiosis of Deep Learning and Differential Equations II.
Speech Enhancement with Score-Based Generative Models in the Complex STFT Domain
Simon Welker, Julius Richter, Timo Gerkmann (2022). "Speech Enhancement with Score-Based Generative Models in the Complex STFT Domain." INTERSPEECH 2022.
Beyond Griffin-Lim: Improved Iterative Phase Retrieval for Speech
Tal Peer, Simon Welker, Timo Gerkmann (2022). "Beyond Griffin-Lim: Improved Iterative Phase Retrieval for Speech." IWAENC 2022.
Speech Enhancement and Dereverberation with Diffusion-based Generative Models
Julius Richter, Simon Welker, Jean-Marie Lemercier, Bunlong Lay, Timo Gerkmann (2022). "Speech Enhancement and Dereverberation with Diffusion-based Generative Models." arXiv preprint arXiv:2208.05830.
Deep Iterative Phase Retrieval for Ptychography
Simon Welker, Tal Peer, Henry N. Chapman, Timo Gerkmann (2022). "Deep Iterative Phase Retrieval for Ptychography." ICASSP 2022.
CMInject: Python framework for the numerical simulation of nanoparticle injection pipelines
Simon Welker, Muhamed Amin, Jochen Küpper (2021). "CMInject: Python framework for the numerical simulation of nanoparticle injection pipelines." Computer Physics Communications.