Michael Baur

Michael Baur

Auto-generated profile, based on publicly available information

Michael Baur

TUM School of Computation Information and Technology Technical University of Munich Germany

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 October 2024 in IEEE Wireless Communications Letters
Verified
Journal Article
Open access
BFBenedikt Fesl
Benedikt Fesl
MBMichael Baur
Michael Baur
FSFlorian Strasser
Florian Strasser
MJMichael Joham
Michael Joham
WUWolfgang Utschick
Wolfgang Utschick
This letter proposes a novel channel estimator based on dm, one of the currently top-rated generative models, with provable convergence to the mse-optimal estimator. A lightweight cnn with positional embedding of the snr information is designed to learn the channel distribution in the sparse angular domain. Combined with an estimation strategy that avoids stochastic resampling and truncates reverse diffusion steps that account for lower snr than the given pilot observation, the resulting dm estimator unifies low complexity and memory overhead. Numerical results exhibit better performance than state-of-the-art estimators.

7 August 2024 in IEEE Transactions on Signal Processing
Verified
Journal Article
Open access
MBMichael Baur
Michael Baur
BFBenedikt Fesl
Benedikt Fesl
WUWolfgang Utschick
Wolfgang Utschick
In this manuscript, we propose to use a variational autoencoder-based framework for parameterizing a conditional linear minimum mean squared error estimator. The variational autoencoder models the underlying unknown data distribution as conditionally Gaussian, yielding the conditional first and second moments of the estimand, given a noisy observation. The derived estimator is shown to approximate the minimum mean squared error estimator by utilizing the variational autoencoder as a generative prior for the estimation problem. We propose three estimator variants that differ in their access to ground-truth data during the training and estimation phases. The proposed estimator variant trained solely on noisy observations is particularly noteworthy as it does not require access to ground-truth data during training or estimation. We conduct a rigorous analysis by bounding the difference between the proposed and the minimum mean squared error estimator, connecting the training objective and the resulting estimation performance. Furthermore, the resulting bound reveals that the proposed estimator entails a bias-variance tradeoff, which is well-known in the estimation literature. As an example application, we portray channel estimation, allowing for a structured covariance matrix parameterization and low-complexity implementation. Nevertheless, the proposed framework is not limited to channel estimation but can be applied to a broad class of estimation problems. Extensive numerical simulations first validate the theoretical analysis of the proposed variational autoencoder-based estimators and then demonstrate excellent estimation performance compared to related classical and machine learning-based state-of-the-art estimators.

7 October 2024 in IEEE Wireless Communications Letters
Verified
Journal Article
Open access
BBBenedikt Böck
Benedikt Böck
MBMichael Baur
Michael Baur
NTNurettin Turan
Nurettin Turan
DSDominik Semmler
Dominik Semmler
WUWolfgang Utschick
Wolfgang Utschick
Prior knowledge about statistical channel characteristics and additional side information, such as estimated parameters shared between radar and communication systems, can both enhance physical layer applications. Generally, the WSSUS property together with the far-field approximation and strongly fluctuating path phases statistically characterize the unconditional wireless channel by a zero mean and Toeplitz-structured covariance matrices. In this letter, we comprehensively categorize side information based on whether the conditioning on this information preserves or abandons these statistical channel features. The established framework combines insights from a generic channel model with representing the channel as a BN. Using our framework, we additionally analyze and improve ML aided channel modeling, clustering and estimation demonstrating its practicality for the physical layer.

19 May 2025 in IEEE Transactions on Machine Learning in Communications and Networking
Verified
Journal Article
Open access
MBMichael Baur
Michael Baur
NTNurettin Turan
Nurettin Turan
SWSimon Wallner
Simon Wallner
WUWolfgang Utschick
Wolfgang Utschick
Generative models are typically evaluated by direct inspection of their generated samples, e.g., by visual inspection in the case of images. Further evaluation metrics like the Fréchet inception distance or maximum mean discrepancy are intricate to interpret and lack physical motivation. These observations make evaluating generative models in the wireless PHY layer non-trivial. This work establishes a framework consisting of evaluation metrics and methods for generative models applied to the wireless PHY layer. The proposed metrics and methods are motivated by wireless applications, facilitating interpretation and understandability for the wireless community. In particular, we propose a spectral efficiency analysis for validating the generated channel norms and a codebook fingerprinting method to validate the generated channel directions. Moreover, we propose an application cross-check to evaluate the generative model’s samples for training machine learning-based models in relevant downstream tasks. Our analysis is based on real-world measurement data and includes the Gaussian mixture model, variational autoencoder, diffusion model, and generative adversarial network. Our results indicate that solely relying on metrics like the maximum mean discrepancy produces inconsistent and uninterpretable evaluation outcomes. In contrast, the proposed metrics and methods exhibit consistent and explainable behavior.

1 February 2022 in IEEE Transactions on Biomedical Engineering
Verified
Journal Article
Closed access
MBMichael Baur
Michael Baur
MRMathias Reisbeck
Mathias Reisbeck
OHOliver Hayden
Oliver Hayden
WUWolfgang Utschick
Wolfgang Utschick
Optical flow cytometry is used as the gold standard in single cell function diagnostics with the drawback of involving high complexity and operator costs. Magnetic flow cytometers try to overcome this problem by replacing optical labeling with magnetic nanoparticles to assign each cell a magnetic fingerprint. This allows operators to obtain rich cell information from a biological sample with minimal sample preparation at near in-vivo conditions in a decentralized environment. A central task in flow cytometry is the determination of cell concentrations and cell parameters, e.g. hydrodynamic diameter. For the acquisition of this information, signal processing is an essential component. Previous approaches mainly focus on the processing of one-cell signals, leaving out superimposed signals originating from cells passing the magnetic sensors in close proximity. In this work, we present a framework for joint cell/particle detection and analysis, which is capable of processing one-cell as well as multi-cell signals. We employ deep learning and compressive sensing in this approach, which involves the minimization of an adaptive norm. We evaluate our method on simulated and experimental signals, the latter being obtained with polymer microparticles. Our results show that the framework is capable of counting cells with a relative error smaller than 2 %. Inference of cell parameters works reliably at both low and high noise levels.

20 November 2024 in IEEE Wireless Communications Letters
Verified
Journal Article
Open access
NTNurettin Turan
Nurettin Turan
MBMichael Baur
Michael Baur
JLJianqing Li
Jianqing Li
WUWolfgang Utschick
Wolfgang Utschick
In this letter, we propose a VQ-VAE-based feedback scheme for robust precoder design in multi-user FDD systems. We demonstrate how the VQ-VAE can be tailored to specific propagation environments, focusing on systems with low pilot overhead, which is crucial in massive MIMO systems. Extensive simulations with real-world measurement data show that the proposed feedback scheme outperforms state-of-the-art AE-based compression schemes and conventional DFT codebook-based schemes. These improvements enable the deployment of systems with fewer feedback bits or pilots.