Peter Prettenhofer retweeted
A few years ago, learning robot learning meant stitching together dozens of papers and courses โ with no clear path from the basics to what state-of-the-art systems actually do.
This was one of the motivations behind creating @ETH's course "Robot Learning: From Fundamentals to Foundation Models", to provide a structured path from first principles all the way to modern foundation models for robotics.
I strongly believe that education should be accessible to everyone, so I have made all lecture recordings publicly available on YouTube.
Creating this course was one of the most challenging projects I have taken on. It was my first time designing and teaching an entire curriculum from scratch, while simultaneously working full-time in industry. On top of that, the course proved to be more popular than expected and we had to scale it to almost 300 students, which was only possible thanks to an amazing team of TAs. Looking back, it was an absolute privilege to teach this class and an incredibly rewarding experience.
If you are getting into robot learning, this is the starting point I wish I had.
๐ Main lectures:
youtube.com/watch?v=X0k14u6pโฆ
๐ค Guest lectures:
youtube.com/watch?v=aG8NPTPhโฆ
๐ Course website: cvg.ethz.ch/lectures/Robot-Lโฆ
Peter Prettenhofer retweeted
๐ฃ ๐ง๐ต๐ฟ๐ถ๐น๐น๐ฒ๐ฑ ๐๐ผ ๐ต๐ผ๐๐ ๐๐ถ๐ฒ๐๐ฒ๐ฟ ๐๐ผ๐
(@fox_dieter17849) ๐ณ๐ผ๐ฟ ๐ฎ ๐ด๐๐ฒ๐๐ ๐น๐ฒ๐ฐ๐๐๐ฟ๐ฒ ๐ฎ๐ ๐๐ผ๐ฑ๐ฎ๐'๐ ๐๐ฒ๐๐๐ถ๐ผ๐ป ๐ผ๐ณ ๐ฅ๐ผ๐ฏ๐ผ๐ ๐๐ฒ๐ฎ๐ฟ๐ป๐ถ๐ป๐ด: ๐๐ฟ๐ผ๐บ ๐๐๐ป๐ฑ๐ฎ๐บ๐ฒ๐ป๐๐ฎ๐น๐ ๐๐ผ ๐๐ผ๐๐ป๐ฑ๐ฎ๐๐ถ๐ผ๐ป ๐ ๐ผ๐ฑ๐ฒ๐น๐ ๐ฎ๐ @ETH!
Today's speaker almost needs no introduction. Dieter is Professor at the @UW and Senior Research Director at @allen_ai, co-author of the landmark textbook ๐ฃ๐ฟ๐ผ๐ฏ๐ฎ๐ฏ๐ถ๐น๐ถ๐๐๐ถ๐ฐ ๐ฅ๐ผ๐ฏ๐ผ๐๐ถ๐ฐ๐, Fellow of @IEEEorg, @RealAAAI, and @TheOfficialACM, and recipient of the ๐ฎ๐ฌ๐ฎ๐ฌ ๐๐๐๐ ๐ฃ๐ถ๐ผ๐ป๐ฒ๐ฒ๐ฟ ๐ถ๐ป ๐ฅ๐ผ๐ฏ๐ผ๐๐ถ๐ฐ๐ ๐ฎ๐ป๐ฑ ๐๐๐๐ผ๐บ๐ฎ๐๐ถ๐ผ๐ป ๐๐๐ฎ๐ฟ๐ฑ and the ๐ฎ๐ฌ๐ฎ๐ฏ ๐๐๐๐๐ ๐๐ผ๐ต๐ป ๐ ๐ฐ๐๐ฎ๐ฟ๐๐ต๐ ๐๐๐ฎ๐ฟ๐ฑ.
Today he will be talking about: "๐ง๐ผ๐๐ฎ๐ฟ๐ฑ ๐๐ผ๐๐ป๐ฑ๐ฎ๐๐ถ๐ผ๐ป ๐ ๐ผ๐ฑ๐ฒ๐น๐ ๐ณ๐ผ๐ฟ ๐๐ผ๐บ๐ฝ๐ฒ๐๐ฒ๐ป๐ ๐ฅ๐ผ๐ฏ๐ผ๐๐" ๐ค
His talk will cover large-scale data generation for robot manipulation, the role of simulation and sim-to-real transfer, and ongoing work at Ai2 on training models that are genuinely competent across a broad set of tasks โ moving beyond narrowly trained, disconnected skills ๐
If you are in Zรผrich and want to join us:
๐ Where: Room NO C 60, Sonneggstrasse 5
๐ When: Today, May 18, 17:00 โ 18:00
See you there!
Peter Prettenhofer retweeted
๐MIT Flow Matching and Diffusion Lecture 2026 Released (diffusion.csail.mit.edu/)!
We just released our new MIT 2026 course on flow matching and diffusion models! We teach the full stack of modern AI image, video, protein generators - theory and practice. We include:
๐บ Videos: Step-by-step derivations.
๐ Notes: Mathematically self-contained lecture notes
๐ป Coding: Hands-on exercises for every component
We fully improved last yearsโ iteration and added new topics: latent spaces, diffusion transformers, building language models with discrete diffusion models.
Everything is available here: diffusion.csail.mit.edu/
A huge thanks to Tommi Jaakkola for his support in making this class possible and Ashay Athalye (MIT SOUL) for the incredible production! Was fun to do this with @RShprints!
#MachineLearning #GenerativeAI #MIT #DiffusionModels #AI
Peter Prettenhofer retweeted
Very pleased by the recent work done by @hug_nicolas to speed-up our pure Python + @numba_jit prototype implementation of gradient boosted trees: github.com/ogrisel/pygbm The master branch is now competitive with LightGBM on the Higgs boson benchmark dataset.
Peter Prettenhofer retweeted
Announcing IMPAC: an IMaging-PsychiAtry Challenge, using data-science to predict autism from brain imaging
paris-saclay-cds.github.io/aโฆ
9000โฌ of prizes to win! More than 2000 individuals scanned!
Organized by @SaclayCDS and @R3RT0's team at @institutpasteur
Peter Prettenhofer retweeted
Donโt compare percentage change on a linear scale; use a log scale instead. -50% (0.5ร) is as big a change as +100% (2ร).
Peter Prettenhofer retweeted
More effective convergence guarantees with AMSGrad, best paper award @ICLR18. Thanks to a slight change in ADAM we can get a stronger update rule for Gradient Descent methods. openreview.net/pdf?id=ryQu7fโฆ #dlearn #Optimization
Peter Prettenhofer retweeted
Tomorrow I'll actually teach my first ever lecture on deep learning, rebuilding all the basic constructs, starting from LDA up to deep rectified networks. Lecture slides available at montefiore.ulg.ac.be/~geurtsโฆ
Peter Prettenhofer retweeted
Worry, if you must, about the future. But you can take comfort in knowing thereโs less of it every day.
Peter Prettenhofer retweeted
. @ContinuumIO we have changed our name to Anaconda, Inc. We are more committed than ever to community: anaconda.com/
Peter Prettenhofer retweeted
I am proud to announce TVM, our new deep learning compiler toolchain news.cs.washington.edu/2017/โฆ tvmlang.org/2017/08/17/tvm-rโฆ @guestrin @uwcse
Peter Prettenhofer retweeted
We just released scikit-learn 0.19. Update with pip or conda. A big thank you to everybody who contributed! scikit-learn.org/0.19/whats_โฆ
Peter Prettenhofer retweeted
Nice example of time-series averaging using soft-DTW tslearn.readthedocs.io/en/laโฆ Part of tslearn by Romain Tavenard
Peter Prettenhofer retweeted
New post: Random-Walk #Bayesian Deep Networks: Dealing with Non-Stationary Data twiecki.github.io/blog/2017/โฆ With animations! #PyMC3
Peter Prettenhofer retweeted
ELI5: debug and explain the behavior of Python Machine Learning models: github.com/TeamHG-Memex/eli5โฆ (sklearn, xgboost, lightning, crfsuite....)
Peter Prettenhofer retweeted
Stochastic online matrix factorization in cython with sklearn compatible API by @arthurmensch: github.com/arthurmensch/modl dict learning & nmf
Peter Prettenhofer retweeted
New post: Native Hadoop file system (HDFS) connectivity in Python wesmckinney.com/blog/python-โฆ #pydata





























