Course type:	Lecture + Exercise
Time:	Lecture: Tuesday, 14:15 - 15:45; Optional exercises: Friday, 10:15 - 11:45
Location:	The course will be in-person. - Weekly flipped classroom sessions will be held on Tuesday in HS 00 026 (G.-Köhler-Allee 101) - Optional exercise sessions will take place on Friday in HS 00 006 (G.-Köhler-Allee 082)
Organizers:	Steven Adriaensen , Abhinav Valada , Mahmoud Safari , Imen Madhi
Web page:	ILIAS - available starting 8am, 20.10.26 (please make sure to also register for all elements of this course module in HISinOne)

Foundations of Deep Learning

Deep learning is one of the fastest growing and most exciting fields. This course will provide you with a clear understanding of the fundamentals of deep learning including the foundations to neural network architectures and learning techniques, and everything in between.

Course Overview

The course will be taught in English and will follow a flipped classroom approach.

Every week there will be:

• a video lecture
• an exercise sheet
• a flipped classroom session (Tuesdays, 14:15 - 15:45)
• an attendance optional exercise session (Fridays, 10:15 - 11:45)

At the end, there will be a written e-exam (an ILIAS test).

Exercises must be completed in groups and must be submitted 2 weeks after their release.
Your submissions will be graded and you will receive weekly feedback.
Your final grade will be solely based on a written examination, however, a passing grade for the exercises is a prerequisite for passing the course.

Course Material: All material will be made available in ILIAS and course participation will not require in-person presence. That being said, we offer ample opportunity for direct interaction with the professors during live Q & A sessions and with our tutors during weekly attendance optional in-class exercise sessions.

Exam: The exam will likely be a test you complete on ILIAS. In-person presence will be required .

Course Schedule

The following are the dates for the flipped classroom sessions (tentative, subject to change):

20.10.26 – Kickoff: Info Course Organization / Team Formation
27.10.26 – Week 1: Intro to Deep Learning
03.11.26 – Week 2: From Logistic Regression to MLPs
10.11.26 – Week 3: Backpropagation
17.11.26 – Week 4: Optimization
24.11.26 – Week 5: Regularization
01.12.26 – Week 6: Convolutional Neural Networks (CNNs)
08.12.26 – Week 7: Recurrent Neural Networks (RNNs)
15.12.26 – Week 8: Attention & Transformers

Winter Break
12.01.26 – Week 9: Practical Methodology
19.01.27 – Week 10: Autoencoders, Variational Autoencoders, GANs
26.01.27 – Week 11: Uncertainty in Deep Learning
02.02.27 – Week 12: AutoML for DL
09.02.27 – Round-up / Exam Q & A

In the first session (on 20.10.26) you will get additional information about the course and get the opportunity to ask general questions (and form groups!). While there is no need to prepare for this first session, we encourage you to already think about forming teams.
The last flipped classroom session will be held on 09.02.26.

Questions?

If you have a question, please post it in the ILIAS forum (so everyone can benefit from the answer) or email dl-orga-ws26@cs.uni-freiburg.de

! Please note we have changed the seminar time from Thursday 3-5pm to 4-6pm !

Tabular data modeling is undergoing rapid transformation driven by foundation models. This seminar is a continuation of our seminar last semester, which examines the most recent advances in our field, with a focus on state-of-the-art approaches across time-series and causal reasoning settings.

Seminar is split into two topic focus tracks:
- Causality (Track a)

- Time Series (Track b)

Kick-off

Link to meeting slides

Pre-registration form

• Thursday 22.10 3pm (Zoom )

Course type:	Seminar
Time	- Kick-off Session: Thursday 22.10 3pm ( Zoom ) - Primer Session 0a (Causality): Thursday 19.11.2026 4pm - Primer Session 0b (Time Series): Thursday 19.11.2026 5pm - Presentation Session 1a (Causality): Thursday 26.11.2026 4pm - Presentation Session 1b (Time Series): Thursday 26.11.2026 5pm - Presentation Session 2a (Causality): Thursday 03.12.2026 4pm - Presentation Session 2b (Time Series): Thursday 03.12.2026 5pm - Presentation Session 3a (Causality): Thursday 17.12.2026 4pm - Presentation Session 3b (Time Series): Thursday 17.12.2026 5pm - Presentation Session 4a (Causality): Thursday 14.01.2027 4pm - Presentation Session 4b (Time Series): Thursday 14.01.2027 5pm - Presentation Session 5a (Causality): Thursday 21.01.2027 4pm - Presentation Session 5b (Time Series): Thursday 21.01.2027 5pm - Presentation Session 6a (Causality): Thursday 28.01.2027 4pm - Presentation Session 6b (Time Series): Thursday 28.01.2027 5pm
Location	SR 00-017/19 in 101
Organizers	Jaris Küken *, Dominika Matus *, Magnus Bühler , Omar Swelam , Jake Robertson , Mustafa Tajjar , Steven Adriaensen and Frank Hutter (*main organizers)
Registration	Open
Language	English

Organization 

• 24 students split into two focus tracks
  • Track a (Causality)
  • Track b (Time Series)
• Each track has 6 papers
  • Each paper is assigned to a group of 2 students
  • One paper is presented in a single session
• Each team gets assigned a buddy team 
  • Dry run presentation with feedback 
• Presentations in pairs
  • Paper presentation (methodology, evaluation, results,...) 
  • Simulate a conference reviewer (critical analysis of the paper) 
  • Additional Evaluations, extensions,... (add-on)

In case of any questions, please contact the main organisers.

Grading

• Presentation (including add-on)
• Discussion
• 1 page add-on report (to be submitted to the supervisor, latest 1 week after the presentation)

Papers

Causality:

1. Paper1(Causality): Causal machine learning for predicting treatment outcomes
2. Paper2(Causality): Causal Foundation Models
3. Paper3(Causality): Causal Foundation Models with Continuous Treatments
4. Paper4(Causality): Causal Longitudinal Prior-Fitted Networks for Counterfactual Outcome Prediction
5. Paper5(Causality): Training and Evaluating Causal Forecasting Models for Time-Series
6. Paper6(Causality): Algorithmic Recourse: from Counterfactual Explanations to Interventions

Time Series:

• Paper1(TS): It's Time
• Paper2(TS): Timer M1
• Paper3(TS): Exploring Representations and Interventions in Time Series Foundation Models
• Paper4(TS): Investigating Compositional Reasoning In Time Series Foundation Models
• Paper5(TS): From Observations to States: Latent Time Series Forecasting
• Paper6(TS): Fractional State Space Transition for Long Sequence Modeling