Topics for graduate theses

Theses are possible at any time in the areas of hardware, simulation and analysis. Some current topics are listed below as examples. For more information and further topics for theses please contact Alexander Kappes (e-mail), IKP room 224.

Current topics for bachelor theses:

• Characterization of a GNN-based noise cleaning algorithm for IceCube-Gen2: Each time a neutrino is detected by IceCube, the photomultiplier tubes (PMTs) measure not only Cherenkov light from the neutrino's secondary charged particles but also noise hits, such as those from radioactive decays in the glass of the pressure vessel of the optical module. Removing this noise is important for the optimal performance of several of IceCube’s neutrino reconstruction algorithms. To address this issue, several noise cleaning algorithms have been developed, and recently, state-of-the-art performance has been achieved using machine learning techniques, specifically Graph Neural Networks (GNNs). This bachelor thesis aims to adapt and evaluate a GNN-based noise cleaning algorithm for the upcoming high-energy extension, IceCube-Gen2. In the cause of this thesis, the student will gain basic knowledge in neural networks with a focus on GNNs, as well as experience in IceCube data processing and high-performance computing.
• Measurment of low charge pulses of photomultipliers: A PMT, upon detecting a photon, yields a charge following a Gaussian-like pattern, known as the "SPE distribution." A precise understanding of this distribution is important for accurate data reconstruction in IceCube. PMTs frequently generate low charge pulses that deviate from the typical Gaussian pattern, but differentiating this deviation from the background distribution remains a challenge. In this bachelor's thesis, the low charge region will be examined comprehensively through diverse measurement techniques.

Current topics for master theses:

• Improving supernova detection in IceCube Upgrade using machine learning: Neutrinos from core-collapse supernovae typically possess energies in the ballpark of tens of MeVs, rendering them impossible to detect using IceCube's standard reconstruction methods. Currently, IceCube identifies potential supernova neutrino bursts by monitoring for a collective increase in the rates across all optical modules. The forthcoming integration of mDOMs in the IceCube Upgrade heralds a promising shift for detecting these events. Given the configuration of 24 photomultipliers per module, the low-energy neutrino events can produce hits across multiple photomultipliers in a single module in a very short time. This master's study will employ machine learning to efficiently discern supernova neutrinos from background noise, focusing on the analysis of hit patterns and correlations across the array of photomultipliers and optical modules.
• Investigation of the performance of real-time reconstruction in IceCube: IceCube has an extensive infrastructure for notifying other experiments in real time when interesting events are detected (real-time alerts). These pipelines are optimized for efficiency and speed and therefore use special reconstruction algorithms. The aim of the program is to find an electromagnetic counterpart to the neutrinos with other telescopes. In this master thesis the performance of the different algorithms will be analyzed, compared and finally improved. In addition, it will be investigated how the uncertainty in the directional reconstruction affects the sensitivity for certain source classes. The work will be carried out in close cooperation with the IceCube group in Bochum, including stays there.
• Seismic Noise Prediction for the Einstein Telescope: This Master's thesis project aims to enhance the Einstein Telescope (ET) Project by developing machine and deep learning algorithms for predicting seismic noise, crucial in gravitational wave analysis. The focus is on forecasting the arrival of seismic waves: P-waves (primary waves that are the fastest and travel through solids, liquids, and gases), S-waves (secondary waves that are slower and only move through solids), and surface waves (which travel along the Earth's surface and cause most of the damage during earthquakes). The candidate will also model the real-time 3D movement of the Earth. This research is key to filtering seismic interference in gravitational wave data. The project aims to significantly enhance the accuracy of gravitational wave detection and contribute to astrophysical research.