Do you think BNNs should be good OOD detectors? We address this question in our talk today at 2:35pm EST at UDL workshop ICML Conference. Also check out our poster at 9:45am EST. Thanks to my awesome colleague Francesco D'Angelo.

I'm happy to announce that my colleague Maria R. Cervera and me published a Python package on how to work with hypernetworks in PyTorch. #hypernet #PyTorch #MachineLearning 🥳 github.com/chrhenning/hyp…

Learning where to learn: Gradient sparsity in meta and continual learning. (arXiv:2110.14402v1 [cs.LG]) ift.tt/3EnwZpD

Ever wondered what makes a neuron spike? Check out experimental + modeling work my wonderful former master students Alex Gastone, Benjamin Ehret & Moritz Buchholz did with me in Kevan Martin's Lab at the Institute of Neuroinformatics at ETH Zürich & UZH Science: biorxiv.org/content/10.110…

I am happy to share that our work on how presynaptic stochasticity can improve energy efficiency and alleviate the stability-plasticity dilemma got accepted at #eLife! The full-text version of our article is fresh of the press: elifesciences.org/articles/69884

It is my great pleasure to announce our latest work from @FritjofHelmchen lab in collaboration with Adriano Aguzzi & Karayannis lab. Many thanks to my brilliant co-authors without their contributions this project could not have been realized in this form: biorxiv.org/cgi/content/sh…

Out now in PLOS Computational Biology! 🥳We combine patch-clamp recordings and modeling to explore how synaptic strength, short-term plasticity, and input synchrony work together to drive neuronal spikes. Enjoy the read! 🙂 journals.plos.org/ploscompbiol/a…

It pays off to persist and push projects over the finish line!🥳🏁 Finally out in PLOS Comp Biol: joint work by the three fantastic master’s students I supervised during my PhD at the Institute of Neuroinformatics ETH Zürich & UZH Science: dx.plos.org/10.1371/journa…

"We showed that transformers (like ChatGPT) can learn on their own to implement algorithms within their architecture," says ETH researcher Johannes von Oswald. 👇 fal.cn/3zXXp #AI #ChatGPT #LargeLanguageModels

I'm very happy to share that my main PhD work is finally out in Nature Neuroscience ! All analysis code and preprocessed data is published online, check it out :) Big thank you to all co authors and everyone who supported me along this long journey! nature.com/articles/s4159…

Uncovering neural activity patterns in the #prefrontal cortex that link sensory stimuli to learned behavioral responses by isolating interpretable activity patterns that are shared among mice performing the same task Benjamin Ehret benjamin grewe nature.com/articles/s4159…