|Course type:||Flipped-classroom Lecture + Exercise|
|Time:||Monday 14:00 - 16:00|
|Location:||SR 01-009/13 (G.-Köhler-Allee 101)|
|Organizers:||Frank Hutter , André Biedenkapp , Heri Rakotoarison ,Neeratyoy Mallik , Eddie Bergman|
Date : April 17, 2023.
Requirement for attending : The Overview lecture from the AutoML MOOC .
Applying machine learning (ML) and in particular deep learning (DL) in practice is a challenging task and requires a lot of expertise. Among other things, the success of ML/DL applications depends on many design decisions, including an appropriate preprocessing of the data, choosing a well-performing machine learning algorithm and tuning its hyperparameters, giving rise to a complex pipeline. Unfortunately, even experts need days, weeks or even months to find well-performing pipelines and can still make mistakes when optimizing their pipelines.The course will discuss meta-algorithmic approaches to automatically search for, and obtain well-performing machine learning systems by means of automated machine learning (AutoML). Such AutoML systems allow for faster development of new ML/DL applications, require far less expert knowledge than doing everything from scratch and often even outperform human developers. In this lecture, you will learn how to use such AutoML systems, to develop your own systems and to understand ideas behind state-of-the-art AutoML approaches.
We strongly recommend that you know the foundations of
- machine learning (ML)
- and deep learning (DL)
We further recommend that you have hands-on experience with:
- Python (3.8+)
- machine learning (scikit-learn)
- deep learning (PyTorch)
The participants should have attended at least one other course for ML and DL in the past.
The lectures are partitioned in several parts, including:
- Hyperparameter Optimization
- Bayesian Optimization for Hyperparameter Optimization
- Neural Architecture Search
- Dynamic Configuration
- Analysis and Interpretability of AutoML
- Algorithm Selection/Meta-Learning
The weekly slides and exercises are available on https://ml.informatik.uni-freiburg.de/github-classroom-automl-2023
The course will be taught in a flipped-classroom style. We will meet weekly once for a combined Q/A session and an exercise. Roughly every week, there will be a new exercise sheet. Most exercises will be practical, involve programming in python and teamwork (teams of up to 3 students!) so that you learn how to apply AutoML in practice.
Lecture/Exercise: Monday 14 - 16
The material is publicly available via the AI-Campus platform . Please register on AI-Campus to access the materials. Grading of the exercises will be done via GitHub classroom.
The links to access the GitHub classroom repositories will be made available on this page . (The password to that page will be announced in the first session).
The course will be taught in English.
The lecture materials are open sourced via https://github.com/automl-edu/AutoMLLecture
Tabular data has long been overlooked by deep learning research, despite being the most common data type in real-world machine learning applications. While deep learning methods excel on many ML applications, tabular data classification problems are still dominated by Gradient-Boosted Decision Trees. More recently, deep learning-based approaches have been proposed which showed remarkable efficiency and performance improvements. In this seminar, we will discuss this recent literature, exploring the most promising techniques and approaches for handling tabular data in deep learning.
|Time||Every Tuesday from 14:15 - 16:00|
|Location||in-person; Room SR 00-006, Building 051|
|Organizers||Herilalaina Rakotoarison , Arbër Zela , Fabio Ferreira , Frank Hutter|
|Web page||Link to the seminar webpage|
|Course type:||Live Lectures + (optional) Exercises|
|Time:||Tuesday 10:00 - 12:00, Friday 10:00 - 12:00|
|Location:||HS 00 036 SCHICK - SAAL (G.-Köhler-Allee 101)|
|Organizers:||Joschka Bödecker , Tim Welschehold , Steven Adriaensen|
Course Content: This course will introduce basic concepts and techniques used within the field of Artificial Intelligence. Among other topics, we will discuss:
- Introduction to and history of Artificial Intelligence
- Problem-solving and search
- Board Games
- Logic and knowledge representation
- Representation of and reasoning with uncertainty
- Machine learning
- Deep Learning
Lectures will roughly follow the book: Artificial Intelligence a Modern Approach (3rd edition)
Organization : There are two lecture slots every week, on Tuesday and Friday (10:15 - 11:45). The first lecture will be on Tuesday 18.04 (see ILIAS for all appointments). During these slots, there will be live (in-person) lectures. These will be in-person only (not hybrid), but recordings will be made available on ILIAS.
Every week, we will also release an exercise sheet (on Friday) to be submitted before next Friday (8:00am). Your submissions will not be graded, but you will receive feedback from our tutors. On Friday, after the live lecture, one of our tutors will present the master solution and answer any questions you may have. The first exercise is due on 28.04. Participation in these weekly sessions and exercises is optional. The only requirement for passing the course is passing the final exam (date/format TBD).