Julian Rossbroich

Image of robots struggling with a social dilemma.

1/ Why does RL struggle with social dilemmas? How can we ensure that AI learns to cooperate rather than compete?

Introducing our new framework: MUPI (Embedded Universal Predictive Intelligence) which provides a theoretical basis for new cooperative solutions in RL.

Preprint🧵👇

(Paper link below.)

1/6 New preprint 🚀 How does the cortex learn to represent things and how they move without reconstructing sensory stimuli? We developed a circuit-centric recurrent predictive learning (RPL) model based on JEPAs.
🔗 doi.org/10.1101/2025...
Led by @atenagm.bsky.social @mshalvagal.bsky.social

Excited to see the paper fully published. It's an important milestone for training SNNs with exact gradients, replacing our earlier tricks of a "delay line augmentation" to capture temporal relationships. Delays can now be learnt alongside weights naturally. Amazing work @mbalazs98.bsky.social !

Spiking neural networks people, this message is for you!

The annual SNUFA workshop is now open for abstract submission (deadline Sept 26) and (free) registration. This year's speakers include Elisabetta Chicca, Jason Eshraghian, Tomoki Fukai, Chengcheng Huang, and... you?

snufa.net/2025/

🤖🧠🧪

What makes visual processing in the brain so powerful and flexible? Very excited to share our new work where we started from SOTA models that accurately predict dynamic brain activity during hours of video watching, and investigated core computations underlying visual perception

There might be a bit of misconception here. What the paper very convincingly shows is that visual cortex does not compute global oddball prediction errors and does not receive any top-down predictions that could be used to compute such prediction errors.

A flexible hippocampal population code for experience relative to reward - Nature Neuroscience Sosa et al. find that hippocampal neural activity in mice encodes both environmental location and experience relative to rewards, spanning distances far from reward, through parallel and flexible popu...

It's officially published!! In my main postdoc work with @markplitt.bsky.social and @lgiocomo.bsky.social, we found that the hippocampus simultaneously encodes an animal's spatial position and its experience relative to reward in parallel population codes. 🧵

www.nature.com/articles/s41...

Memory by a thousand rules: Automated discovery of functional multi-type plasticity rules reveals variety & degeneracy at the heart of learning Synaptic plasticity is the basis of learning and memory, but the link between synaptic changes and neural function remains elusive. Here, we used automated search algorithms to obtain thousands of str...

We just pushed “Memory by a 1000 rules” onto bioRxiv, where we use clever #ML to find #plasticity quadruplets (EE, EI, IE, II) that learn basic stability in spiking nets. Why is it cool? We find 1000s!! of solutions, and they don’t just stabilise. They #memorise! www.biorxiv.org/content/10.1...

Dynamic basal ganglia output signals license and suppress forelimb movements - Nature Basal ganglia output neurons fire dynamically in bidirectional and movement-specific patterns to license forelimb movements. 

Forelimb movement control at the basal ganglia - brainstem interface!

Happy to finally share this work from me and @harsh-kanodia.bsky.social with Silvia Arber!

@biozentrum.unibas.ch @fmiscience.bsky.social

www.nature.com/articles/s41...

New #NeuroAI #compneurosky preprint! To better understand how target-directed learning works in the brain, we sought to engineer an artificial neural network capable of solving complex image classification tasks that comprises only experimentally-supported biological building blocks. (1/15)

I've spent much of my PhD thinking about E/I balance, and our latest preprint represents the culmination of that journey. Huge thanks to @fzenke.bsky.social for guiding me.

Looking forward to your thoughts & comments.

Oja's plasticity rule overcomes several challenges of training neural networks under biological constraints Deep neural networks have achieved impressive performance through carefully engineered training strategies. Nonetheless, such methods lack parallels in biological neural circuits, relying heavily on n...

New preprint with my postdoc, Navid Shervani-Tabar, and former postdoc, Marzieh Alireza Mirhoseini.

Oja’s plasticity rule overcomes challenges of training neural networks under biological constraints.

arxiv.org/abs/2408.08408

Top-down feedback is ubiquitous in the brain and computationally distinct, but rarely modeled in deep neural networks. What happens when a DNN has biologically-inspired top-down feedback? 🧠📈

Our new paper explores this: elifesciences.org/reviewed-pre...

How does our brain predict the future? Our review of predictive processing + research program is now on arXiv arxiv.org/abs/2504.09614
50+ neuroscientists distributed across the world worked together to create this unique community project.

Check out our latest paper today in Nature: “Goal specific hippocampal inhibition gates learning” www.nature.com/articles/s41...
By Nuri Jeong, Xiao Zheng, Abby Paulson, Steph Prince and colleagues.

1/7: Super excited to share our new paper! This one should be of interest to neuroscientists and deep learning theory folks. This paper was a collaboration with Alexandre Payeur, @averyryoo.bsky.social, Thomas Jiralerspong, @mattperich.bsky.social, Luca Mazzucato, @glajoie.bsky.social

Such wonderful work! Congrats Emerson and @neuronaud.bsky.social 🪅

A prospective code for value in the serotonin system - Nature Merging ideas from reinforcement learning theory with recent insights into the filtering properties of the dorsal raphe nucleus, a unifying perspective is found explaining why serotonin neurons are ac...

I'm excited to share that the last chapter of my PhD thesis is now published in Nature! 🍾

What drives serotonin neurons? We think it's the expectation of future reward and --- critically --- how fast this expectation is increasing. 📈

doi.org/10.1038/s415...

1/6