|
|
Einrichtungen >> Fakultät Wirtschaftsinformatik / Angewandte Informatik >> Bereich Angewandte Informatik >> Lehrstuhl für Erklärbares Maschinelles Lernen >>
|
xAI-Proj-B: Bachelorprojekt Erklärbares Maschinelles Lernen- Dozentinnen/Dozenten
- Jonas Alle, Prof. Dr. Christian Ledig
- Angaben
- [PJS]
Rein Präsenz
4,00 SWS, Unterrichtssprache Englisch
Zeit und Ort: Do 14:00 - 18:00, WE5/04.003 (außer Do 6.2.2025)
- Voraussetzungen / Organisatorisches
- Interest and registration
Registration via central registration or email to jonas.alle@uni-bamberg.de with matriculation number, degree program, and completed ML-related modules before October 18, 2024.
Eligibility
B.Sc. KI & Data Science, B.Sc. AI, [B.Sc. WI (only after prior consultation with the examination office), M.Sc. ISoSySc (only possible with Learning Agreement)]
Requirements
None
VC-Course: https://vc.uni-bamberg.de/course/view.php?id=70918
Beneficiaries
Critical thinking, Knowledge in programming (Python), Hands-on knowledge in machine learning, scientific writing, LaTeX
- Inhalt
- Topic:
Data Comprehension and Visualization
Overview:
In this project you will get a hands-on experience to the fields of data engineering, data analysis, and machine learning. Given a fine-grained dataset - that is, a dataset which provides next to typical class information also some detailed attribute occurrences per image sample - you will have the space and opportunity to creatively experiment and explore the data in a fully interest-driven manner. We will use the CUB dataset containing 200 different bird species in a total of almost 12k images each of which has further annotations of bird "parts" (beak, belly, wing, leg, ...) and attribute groups (part color, part shape, part pattern, ...).
The project is meant as an open challenge to find underlying structure, e.g., correlations, biases, symmetries, variances, etc., across the different annotation levels by using common data engineering and analysis tools like density estimation and dimensionality reduction. You will generate insightful visualizations and as a group put together a big-picture understanding of the dataset.
Following this, you will test different methods to classify the data. Both machine learning (e.g. kNN, decision trees) as well as deep learning (e.g. CNNs) approaches are possible! You are free to explore whatever method/data aspect interests you the most!
Prior machine learning knowledge is not needed but you should bring a descent amount of curiosity and readiness to learn about new algorithms independently. Experience in the programming language Python is an advantage (especially with libraries like numpy and scikit-learn but you will also have the time to pick it up along the way.
Initial meeting: 17/10/24 (General info, Q&A and preliminary overview - not mandatory)
Kick-off: 24/10/24 (mandatory for participants)
- Englischsprachige Informationen:
- Title:
- xAI-Proj-B: Bachelor Project Explainable Machine Learning
- Credits: 6
- Zusätzliche Informationen
- Erwartete Teilnehmerzahl: 15
Hinweis für Web-Redakteure:
Wenn Sie auf Ihren Webseiten einen Link zu dieser Lehrveranstaltung setzen möchten, verwenden Sie bitte einen der folgenden Links:Link zur eigenständigen Verwendung Link zur Verwendung in Typo3
|
 |
|
|
UnivIS ist ein Produkt der Config eG, Buckenhof |
|
|