|
Size: 1747
Comment:
|
Size: 1420
Comment:
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 8: | Line 8: |
| || Credits || 3 ECTS, Elective in the modules "Machine Learning I", "Machine Learning II", and "Cognitive Algorithms" || | |
| Line 15: | Line 14: |
| * Federated Learning and Federated Data Preparation. | * Federated/Deep Ensemble Learning and Data Management Systems. |
| Line 17: | Line 16: |
| * Compression of ML Systems: From pruning through knowledge distillation to quantization. * Learning on multi-modal data, e.g. DALLE. * ML and Data Management Systems for training very large generative models, e.g. GPT. |
* Compression of ML Systems. |
| Line 23: | Line 20: |
| * Scalable Bayesian Learning. * MLOps: Building ML Training pipelines for data preparation, model training, debugging, versioning, and monitoring. |
* Building ML pipelines for large large scale data preparation, model training and model debugging, versioning, and monitoring. |
Joint Seminar on Machine Learning and Data Management Systems
Date |
First meeting: Tuesday 01.06.2022, 14-15, Presentation: TBA |
Room |
First meeting: Via Zoom (TBA), Presentation: MAR4.033 |
Trainers |
Matthias Böhm, Dennis Grinwald |
Contact |
This is a joint research-oriented seminar of the Machine Learning Group and the Data Management Group. Throughout the seminar, students will have the opportunity to learn about recent advances in the intersection of Machine Learning and Data Management Systems.
Interested students are required to participate in the kick-off meeting after which they will select, read, understand, and (if possible) programmatically evaluate one of the eligible papers (TBA), before giving a final 10-15 min presentation in the English language at the end of the semester. More details will be discussed during the Kick-off meeting.
Example topics include:
- Federated/Deep Ensemble Learning and Data Management Systems.
- Carbon-aware data management systems and Machine Learning.
- Compression of ML Systems.
- Continual, Lifelong, and Online Learning.
- Data management systems for Lifelong Learning.
- Hashing and sketches.
- Building ML pipelines for large large scale data preparation, model training and model debugging, versioning, and monitoring.
- AutoML.