Semester
Überblick
- Das Seminar ist auf ca. 60 Teilnehmer/innen des Masterstudiengangs Data Science begrenzt.
- Um an dem Seminar teilzunehmen, müssen Sie sich im AlmaWeb für das Modul Skalierbare Datenbanktechnologien 1 (10-INF-DS01) und das Seminar anmelden sowie an der Einführungsveranstaltung teilnehmen, bei der die endgültige Platz- und Themenvergabe erfolgt.
- Bei Fragen und Problemen zur An- und Abmeldung wenden Sie sich bitte an das Studienbüro via einschreibung [at] math.uni-leipzig.de (einschreibung(at)math.uni-leipzig.de)
- Die studentischen Vorträge finden an den Freitagsterminen im Jan./Feb 2027 (jeweils ab 13:15 Uhr) statt.
- Alle Materialien und Informationen finden Sie im entsprechenden Moodle-Kurs.
- Einführungsveranstaltung: 23.10.2026, 13:30 Uhr, HS 10
Leistungsbewertung
Ein erfolgreiches Seminar setzt die Teilnahme an allen Seminarterminen voraus, die selbständige Erarbeitung eines Themas sowie einen Vortrag über das Thema.
Literatur
Für das Verständnis ist es hilfreich sich folgende Studentenarbeiten oder Originalliteratur Quellen zum Thema Deep Learning anzuschauen.
| Thema | Quelle |
|---|---|
| Einleitung Deep-Learning | Studentenarbeit,[1],[2] |
| Autoencoder und CNN | Studentenarbeit,[1] |
| Recurrent Neural Networks | [1] [2] |
| How to read a paper | [1] |
| How to avoid boring presentations (EDBT 2024- Tutorial) | [1] |
Themen
>>Kontakt-Formular für Studierende, um ihren Betreuenden eine alternative Mail-Adresse mitzuteilen<<
| P1 | Toward Privacy-Preserving LLM Personalization | Lange | [1] |
| P2 | Continual Unlearning with Preservation of LLM Knowledge | Lange |
|
| LLM1 | Trust Regions in Reinforcement Learning for LLMs | Lange | [1] |
| LLM2 | Memorization vs. Generalization in LLMs | Augenstein | [1], [2] |
| LLM3 | Hallucination in LLMs: Detection & Mitigation | Augenstein | [1] |
| LLM4 | Efficient Fine-Tuning with LoRA | Augenstein | [1] |
| LLM Agents & Tool Use: ReAct and Toolformer | Augenstein | [1],[2] | |
| LLM6 | Multi-Agent LLM Systems: MetaGPT & CAMEL | Augenstein | [1],[2] |
| Advanced Retrieval: Hypothetical Document Embeddings (HyDE) | Augenstein | [1] | |
| LLM8 (Backup) | Matryoshka Embeddings: Flexible-Dimension Representations | Augenstein | [1],[2] |
| LLM9 (Backup) | Domain-Adaptation of Embedding Models | Augenstein | [1] |
| LLM10 (Backup) | Model Compression: Activation-aware Weight Quantization (AWQ) | Augenstein | [1] |
| RAG1 | Agentic Retrieval-Augmented Generation | Augenstein | [1],[2] |
| RAG2 | Evaluating RAG Systems: RAGAs & RAGChecker | Augenstein | [1],[2] |
| Graph-based Retrieval: HippoRAG & LightRAG | Augenstein | [1],[2] | |
| Cache-Augmented Generation: An Alternative to RAG | Augenstein | [1] | |
| LLM-Compiled Wikis as a RAG Alternative | Augenstein | [1],[2] | |
| Context Utilisation for Retrieval-Augmented Generation | Christen | [1] | |
| Data Management | |||
| DM1 | Table Discovery | Christen | [1],[2] |
| DM2 | Knowledge Graph Entity Alignment | Christen | [1] |
| DM3 | Geospatial entity resolution | Rohde | [1] |
| DM4 | Vehicle Re-Identification and Tracking: Algorithmic Approach, Challenges and Future Directions | Rohde | [1] |
| DM5 | egenioussBench: A New Dataset for Geospatial Visual Localisation | Rohde | [1] |
| DM6 | End-to-End Framework for the Automatic Matching of Omnidirectional Street Images and Building Data and the Creation of 3D Building Models | Rohde | [1] |
| Time Analysis & Temporal Graphs | |||
| G1 | CLaP - State Detection from Time Series | Ammar | [1] |
| G2 | Traffic congestion propagation inference using dynamic Bayesian graph convolution network | Ammar | [1] |
| G3 | Joint learning of tensorial incompact multi-view unsupervised feature and missing view impuation | Parvaneh | [1] |
| G4 | Topological Autoencoders | Parvaneh | [1] |
| PH1 | PINN-Sed: Interconnected Physics-Informed Machine Learning Model for High Resolution Suspended Sediment Transport in River Network | Uhrich | [1] |
| Curvature-aware optimization for high-accuracy physics-informed neural networks | Uhrich | [1] | |
| PH3 | Physics-informed graph transformer network for predicting cable-stayed bridge structural deflection response | Uhrich | [1] |
| Biomedical Applications | |||
| BM1 | nnU-Net: fully automated biomedical image segmentation | Martin | [1] |
| BM2 | 3DINO-ViT: Self-Supervised Learning method for medical imaging | Martin | [1] |
| BM3 | Predicting brain tumor malignancy in CLE imaging | Martin | [1] |
| BM4 | Mamba: Linear-Time Sequence Modeling with Selective State Spaces | Neumann | [1] |
| BM5 | MedFlowSeg: Flow Matching for Medical Image Segmentation with Frequency-Aware Attention | Neumann | [1] |