Dominik Dold

Drowning Doesn’t Look Like Drowning Drowning is not the violent, splashing call for help that most people expect.

Thanks for the reminder @janemunday.bsky.social. Every summer, I repost this article DROWNING DOES NOT LOOK LIKE DROWNING. To date, I know of FOUR kids who were saved after someone who'd clicked on the link learnt how to spot actual drowning. Take time to read and pass on.

slate.com/technology/2...

This work has been funded by @ec.europa.eu and was performed at @univie.ac.at 🤗

Causal pieces: analysing and improving spiking neural networks piece by piece We introduce a novel concept for spiking neural networks (SNNs) derived from the idea of "linear pieces" used to analyse the expressiveness and trainability of artificial neural networks (ANNs). We pr...

You can find the preprint on arXiv: arxiv.org/abs/2504.14015

We believe that this measure can be used to study and improve various aspects of spiking neural networks, from neuron models to initialisation schemes and training methods! 😊

2. The more causal pieces the training data falls into before training, the higher the chances that the network trains successfully and reaches a high performance 📈 Hence, this measure can be used to guide initialisation of spiking neural networks.

We found that the number of such causal pieces has some cool properties:

1. The approximation error is lower bounded by an expression depending on the inverse squared of the number of causal pieces. More pieces, less error (which does not mean better generalization though)!

A causal piece is, quite literally, a piece of the input (and parameter) space where the network output is always caused by the same network components. Or simply put: the path through the network stays the same.

That's all the differently coloured regions shown above - one colour 🟩 = 🧩 one piece!

In spiking neural networks, neurons communicate - as in the brain - via short electrical pulses⚡(spikes). But how can we formally quantify the (dis)advantages of using spikes? 🤔

In our new preprint, @pc-pet.bsky.social and I introduce the concept of "Causal Pieces" to approach this question!

I'd say that's pretty good loot ;)

So I guess you didn't find two huge guys with pumpkins on their head down there? :D

I (unfortunately) have no book recommendations, but when there: definitely check out the Herculaneum archaeological site, absolutely mind-blowing! :)

If you are interested in machine learning, #astronomy, (or both!), have a look at this #hackathon hosted by #ESA in Madrid, 16-17 Jan 2025!

www.ariel-datachallenge.space/esa-datalabs...

Early career researchers can apply for funding :)

For the Blueskyers interested in #NeuroAI 🧠🤖,
I created a starter pack! Please comment on this if you are not on the list and working in this field 🙂

go.bsky.app/CscFTAr

🙋🏼