Anna Levina

Mainly science = computational neuroscience, self-organization, critical dynamics, plasticities. Sometimes politics. Opinion’s are own. RT ≠ endorse. She/her

[Anna Levina @SelfOrgAnna]

We are hiring a Postdoc in Computational Neuroscience and ML at the University of Tübingen. Work with me and Georg @GMartius on expressive neurons, recurrent networks, and neuro-inspired AI for RL and robotics.
https://t.co/26PkX701Zm the related work:
https://t.co/Gkl8frUqCt

[Aaron Spieler @AaronSpieler]

I'm thrilled 😊 to announce that our paper on the Expressive Leaky Memory (ELM) neuron got accepted to #ICLR2024 🎉 Huge thanks to @nasim_rahaman @GMartius @bschoelkopf @SelfOrgAnna 🙏!

OpenReview: https://t.co/cI7gWVrH1R
TweePrint: https://t.co/VEIm2hpivp

[Aaron Spieler @AaronSpieler]

1) I'm excited to present "Scaling Laws and Tradeoffs in Recurrent Network of Expressive Neurons" 🎉, where we question the optimality of simple neurons for network performance, and capture scaling behaviours using a minimal information theoretic model: https://t.co/g2Co7752xi

[Anna Levina @SelfOrgAnna]

@codewithimanshu @GMartius No, we are trying to identify and implement important mechanisms, not to duplicate the brain, it will remain a model ;)

[Anna Levina @SelfOrgAnna]

We are hiring a Postdoc in Computational Neuroscience and ML at the University of Tübingen. Work with me and Georg @GMartius on expressive neurons, recurrent networks, and neuro-inspired AI for RL and robotics.
https://t.co/26PkX701Zm the related work:
https://t.co/Gkl8frUqCt

[Aaron Spieler @AaronSpieler]

I'm thrilled 😊 to announce that our paper on the Expressive Leaky Memory (ELM) neuron got accepted to #ICLR2024 🎉 Huge thanks to @nasim_rahaman @GMartius @bschoelkopf @SelfOrgAnna 🙏!

OpenReview: https://t.co/cI7gWVrH1R
TweePrint: https://t.co/VEIm2hpivp

[Anna Levina @SelfOrgAnna]

How to better represent 3D rotations? It was a very fun project that involved a lot of learning for me. Now accepted to #ICML2024 . Congratulations to the whole team!

[Georg Martius @GMartius]

Training neural networks on 3D rotations feels at times like alchemy. Which rotation representation should we use? How critical is the choice of loss? Our recent preprint looks at these questions with clear visuals and simple math. Check out 👇 https://t.co/75MetpkHu6
1/ 🧵

[TTchumatchenko @TTchumatchenko]

📢 22-25 April 2024 Digital workshop on Dendritic computation, synaptic plasticity & molecular mechanisms. Join @cian_neuro and me in learning about the recent findings of friends and colleagues working in this exciting field. https://t.co/Y0hQmj7oNo

[Anna Levina @SelfOrgAnna]

In case you were just shy to ask - I am happy to welcome MSC-applicants and support applications centered on relevant topics (self-organization in neuronal networks, E/I balance in recurrent networks, plasticity, criticality, network dynamics (and more)...)

[Roxana Zeraati @roxana_zeraati]

Exited to give a talk at #Cosyne2024 about our recent work on how different learning curricula give rise to distinct mechanisms for performing long-memory tasks.

If you’re interested drop by session 8 on Saturday.

Live stream: https://t.co/3mDCghjHCO

https://t.co/4QWAyUfVhO

[Anna Levina @SelfOrgAnna]

Look at the cool tweeprint by @roxana_zeraati about our paper assessing the role of individual and network-mediated timescales in memory-dependent computational tasks. Proud of the whole lab participating in this paper (to be presented at #ICLR )

[Roxana Zeraati @roxana_zeraati]

Thrilled to share our #ICLR2024 paper (co-led w/ @sinak12345 and great collaboration w/@m_giannakakis @schaefertim1 @GMartius @SelfOrgAnna).

We show that mechanisms of generating long timescales for memory tasks depend on how RNNs learn the task.
🧵👇

https://t.co/rYjIPwnsRc

[Roxana Zeraati @roxana_zeraati]

Thrilled to share our #ICLR2024 paper (co-led w/ @sinak12345 and great collaboration w/@m_giannakakis @schaefertim1 @GMartius @SelfOrgAnna).

We show that mechanisms of generating long timescales for memory tasks depend on how RNNs learn the task.
🧵👇

https://t.co/rYjIPwnsRc

[TTchumatchenko @TTchumatchenko]

My group and I are looking forward to welcome a new PhD student in our midst. Join us and help us model dendritic proteins in health and disease.

[Intelligent Systems @MPI_IS]

Join our IMPRS-IS PhD program - a partnership with @uni_stuttgart & @uni_tue! Application deadline Nov 15, 2023. Apply at https://t.co/veVpmjl215

[Eleonora Russo @russo_eleon]

Interested in investigating how neuronal dynamics supports the execution of cognitive functions?
Join us! 📢PhD position in the Brain Dynamics Laboratory at @SantAnnaPisa, Italy.
#manifold #RNN #compneuro

Deadline for the application: 4 August.
https://t.co/G7Y5aDrEtB

[Dmitry Kobak @hippopedoid]

Very impressive joint investigation by @meduza_en and @mediazona_en on MALE excess mortality in Russia during the war:
* https://t.co/Ugcql8EFrj
* https://t.co/je0EwCclYo
Spoiler: ~25K in 2022, ~50K until now.

Here is the little part that I did, based on the Rosstat data. [1/5]

[Anna Levina @SelfOrgAnna]

@KordingLab Here is now also a tweeprint (this is a bit embarrassing, but we manned the print later than your comment ;)

[Aaron Spieler @AaronSpieler]

1) We present our Expressive Leaky Memory (ELM) neuron 🧠, which can capture detailed cortical neuron I/O with few thousand params (1000x improv 🤯) but can also solve complex long-range tasks such as Pathfinder-X (16k context len 🍀)!

Preprint: https://t.co/dEdWliUsK9

[Aaron Spieler @AaronSpieler]

1) We present our Expressive Leaky Memory (ELM) neuron 🧠, which can capture detailed cortical neuron I/O with few thousand params (1000x improv 🤯) but can also solve complex long-range tasks such as Pathfinder-X (16k context len 🍀)!

Preprint: https://t.co/dEdWliUsK9

[Kording Lab 🦖 @KordingLab]

Single neurons can in principle be super powerful: https://t.co/bVm6p7cg8A

[Shervin Safavi @neuroprinciples]

1/12 Happy to share our brief and sweet results on signatures of criticality in efficient coding networks!

Preprint: https://t.co/VS41Qbi2oa
Talk: https://t.co/GkpWh9hPCX
#tweeprint 👇

[Anna Levina @SelfOrgAnna]

Knowing line-up and organisers, this will be an exciting exchange

[Johannes Zierenberg @jozierenberg]

Excited about the interface between stat. phys. and machine learning? Apply for our workshop and discuss with J. Byers, @marylougab, @sebastiangoldt, @InackEstelle, @kadmonj, S. Karpitschka, M. Klopotek, B. Seoane, @wangleiphy, M. Wright, R. Zecchina, ... https://t.co/EBB6MUuriz

[Jakob Macke @jakhmack]

Come study #MachineLearning in Tübingen: 10 days left to apply to our International MSc Program! Tübingen boasts an amazing research environment #AI/#ML, with several world-class institutions @Cyber_Valley @uni_tue Tuebingen AI @ml4science @MPI_IS, https://t.co/IU0dW7sz2b

[Tatiana Engel @EngelTatiana]

What approach can explain the link between the brain and behavior? Shall we focus on neural circuits or manifolds? With @chrismlangdon & @MGENK, we provide a unifying perspective on manifolds and circuits, just out in @NatRevNeurosci: https://t.co/sM4xKUEZ6o. #tweeprint 👇