Attachment
Author
Date
Last edited time
Nov 23, 2022 12:35 PM
URL
Format
- What: An in-person, hands-on, free workshop
- For whom: the workshop is limited to PhD students, postdocs, and PIs affiliated with the University of Tübingen
- When: March 30th, 2022, from 9:00 to 13:00 (with pauses)
- Where: Tübingen AI Center, Maria-von-Linden-str. 6, 72076 Tübingen, Lecture Hall (ground floor, Room 00-28/A7)
- Instructors: and
Álvaro Tejero-Cantero, ML ⇌ Science Colaboratory, Cluster of Excellence ML for ScienceSeth Axen - Covid regulations: 3G-rules apply (proof of vaccination or past infections or negative test). Wearing a FFP2 mask is mandatory for participants.
Content
In the tutorial part, we will discuss with the aid of interactive demonstrations topics such as
- Foundational ML notions: data features, model complexity, generalization, evaluation of performance, ...
- Neural networks: automatic feature extraction, choice of architecture, iterative training, pretraining and fine-tuning.
- Data types: text, time series, relational data, images, ...
- Programming ML models vs. traditional software.
- The ML workflow in practice, from data collection to performance evaluation.
- Some limitations of ML for science.
In the workshop part, we will jointly look at specific use-cases brought by the attendees and discuss how to design a ML project for each of them.
Prerequisites
- Scholars from all disciplines are invited to join; no math or programming skills will be assumed.
- Bring your own laptop, and please check your eduroam config before coming!
- In connection to one of your research problems, think about the characteristics of relevant datasets that you have access to (how many records, what fields? ...). Be ready to discuss your use-case with the group; we will provide suggestions to help you make a practical machine learning project out of it.
Signing up
Apply by filling the form. You will hear from us approximately one week before the event.
@March 18, 2022 registrations are now closed. We received many more applications than expected! We will offer the workshop again later in the year, and are considering adding workshops on specific techniques of ML for those of you with specialized needs.
Thank you for your interest in our event!
Questions?
Write to us at mlcolab@inf.uni-tuebingen.de.
