Michalis Pagkalos

Have you tried building neuronal models with #DendroTweaks? It’s a Python toolbox we made to help you visualize and customize single-neuron models with active dendrites. And it's just got a major update! ✨ Here's what's new: 🧵

@yiotapoirazi.bsky.social

🔍 Find out more about Dendrify 2.2 and dive into the new features here: dendrify.readthedocs.io/en/latest/ 🌐

For updates, follow me here: bsky.app/profile/mrde... 🚀

And make sure to check out our lab's page: bsky.app/profile/dend... 🔬

Thanks for your support and stay tuned for more! 👀

📚 Documentation improvements!

We’ve added new tutorials, examples, and reference docs. Plus, numerous aesthetic & language enhancements to make your experience smoother. See the before/after here: (4/5)

🔧 Model Customization is here!

Now, you can fully customize the built-in models in Dendrify. Add new equations, replace existing ones, and introduce new parameters to tailor your simulations to your exact needs!
Find out more here: dendrify.readthedocs.io/en/latest/tu... (3/5)

New feature: ✨dSpike Playground✨

We've introduced a graphical tool for interactively visualizing, exploring, and calibrating dendritic spikes in Dendrify. AKA dSpikes finally made easy! (2/5)

🚨BREAKING🚨
Dendrify 2.2 is live and it’s one of our biggest updates yet . This release focuses on user experience and customisability, all built from your feedback! 📝

💡 What’s new?
🔹 dSpike Playground
🔹 Model Customization
🔹 Improved Docs

@yiotapoirazi.bsky.social

See 🧵👇 (1/5)

Dendrites endow artificial neural networks with accurate, robust and parameter-efficient learning - Nature Communications Artificial neural networks, central to deep learning, are powerful but energy-consuming and prone to overfitting. The authors propose a network design inspired by biological dendrites, which offe...

Towards a More Sustainable Future for AI:
Led by Dr. Chavlis, our research, published in @naturecomms.bsky.social, demonstrates how incorporating dendritic features into Artificial Neural Networks (ANNs) can significantly improve their efficiency and performance.
www.nature.com/articles/s41...