The SCCS Colloquium is a forum giving students, guests, and members of the chair the opportunity to present their research insights, results, and challenges. Do you need ideas for your thesis topic? Do you want to meet your potential supervisor? Do you want to discuss your research with a diverse group of researchers, rehearse your conference talk, or simply cheer for your colleagues? Then this is the right place for you (and you are also welcome to bring your friends along).
Upcoming talks
Mehmet Rıfat Özkurt: Extensions to the Deep Learning Methods on Full Waveform Inversion
SCCS Colloquium |
Full waveform inversion (FWI) reconstructs the material properties of a medium by iteratively fitting simulated wave signals to measured ones. Applied to ultrasonic non-destructive testing (NDT), it images the interior of a physical specimen from signals collected at its surface and resolves defect geometry more accurately than established array imaging methods. However, the underlying inverse problem for reconstruction is ill-posed, the iterative solution is computationally expensive, and the reconstruction is highly dependent on the initial guess. Recent work addresses these issues with deep learning, most notably by using a convolutional neural network (CNN) as the ansatz for the material field, pretrained on synthetic specimens and fine-tuned online through the inversion itself. However, what the pretraining contributes, and whether other representations or network architectures would serve this problem better remain open questions with varying views in the literature. This work continues that line of research and investigates the open questions in three experiments. First, we conduct an ablation study where we freeze and re-initialize parts of the pretrained U-Net during the fine-tuning inversion and show that the benefit of transfer learning lies in the learned features of its encoder, as opposed to a favorable initialization of the weights alone. Second, we adapt four INR-based methods from continuous-representation FWI (CR-FWI) in the seismic imaging literature to the NDT setting. These converge far more slowly than the U-Net and struggle to reconstruct the sharp boundaries of voids, locating compact defects with a Dice coefficient on par with the U-Net (up to 0.98), while failing on thinner cracks. Finally, we replace the U-Net ansatz by a hierarchical vision transformer (SegFormer), with and without ImageNet initialization, and compare the two architectures over pretraining sets of 250 to 15,000 specimens. The U-Net remains the more accurate network after the inversion, but the SegFormer predicts the defect from the first FWI gradient better than the U-Net, at almost every pretraining size when initialized from ImageNet, and from 5,000 samples onwards for the randomly initialized version. The ImageNet-initialized SegFormer reaches a Dice coefficient of 0.86 against the U-Net’s 0.78 at 15,000 pretraining samples, in the initial prediction before the online inversion. This stronger initial prediction makes the SegFormer a promising candidate both for surpassing the convolutional network through further work on its online fine-tuning, and for direct defect prediction where the inversion loop is too expensive to run.
Master's thesis presentation. Mehmet is advised by Qing Sun.
You don't want to miss a talk? Subscribe to our mailing list and our Colloquium calendar .
Contribute a talk
To register and schedule a talk, you should fill the form Colloquium Registration at least four weeks before the earliest preferred date, at the end of a semester typically even earlier reservation is necessary due to a very high influx of presentations. Keep in mind that we only have limited slots, so please plan your presentation early! In special cases, contact colloquium@mailsccs.in.tum.de.
Colloquium sessions are now on-campus. We have booked room MI 00.13.054 for WS26/27. You can either bring your own laptop or send us the slides as a PDF ahead of time. The projector only has an HDMI connection, so please bring your own adapters if necessary.
We invite students doing their Bachelor's or Master's thesis, as well as IDP, Guided Research, or similar projects at SCCS to give one 20min presentation to discuss their results and potential future work. The time for this is typically after submitting your final text. Check also with your study program regarding any requirements for a final presentation of your project work.
New: In regular times, we will now have slots for presenting early stage projects (talk time 2-10min). This is an optional opportunity for getting additional feedback early and there is no strict timeline.
Apart from students, we also welcome doctoral candidates and guests to present their projects.
During the colloquium, things usually go as follows:
- 10min before the colloquium starts, the speakers setup their equipment with the help of the moderator. The moderator currently is Ana Cukarska. Make sure to be using an easily identifiable name in the online session's waiting room.
- The colloquium starts with an introduction to the agenda and the moderator asks the speaker's advisor/host to put the talk into context.
- Your talk starts. The scheduled time for your talk is normally 20min with additional 5-10min for discussion.
- During the discussion session, the audience can ask questions, which are meant for clarification or for putting the talk into context. The audience can also ask questions in the chat.
- Congratulations! Your talk is over and it's now time to celebrate! Have you already tried the parabolic slides that bring you from the third floor to the Magistrale?
Do you remember a talk that made you feel very happy for attending? Do you also remember a talk that confused you? What made these two experiences different?
Here are a few things to check if you want to improve your presentation:
- What is the main idea that you want people to remember after your presentation? Do you make it crystal-clear? How quickly are you arriving to it?
- Which aspects of your work can you cover in the given time frame, with a reasonable pace and good depth?
- What can you leave out (but maybe have as back-up slides) to not confuse or overwhelm the audience?
- How are you investing the crucial first two minutes of your presentation?
- How much content do you have on your slides? Is all of it important? Will the audience know which part of a slide to look at? Will somebody from the last row be able to read the content? Will somebody with limited experience in your field have time to understand what is going on?
- Are the figures clear? Are you explaining the axes or any other features clearly?
In any case, make sure to start preparing your talk early enough so that you can potentially discuss it, rehearse it, and improve it.
Here are a few good videos to find out more:
- Simon Peyton Jones: How to Give a Great Research Talk (see also How to Write a Great Research Paper)
- Susan McConnell: Designing effective scientific presentations
- Jens Weller: Presenting Code
Did you know that the TUM English Writing Center can also help you with writing good slides?
Work with us!
Do your thesis/student project in Informatics / Mathematics / Physics: Student Projects at the SCCS.