Anna Vettoruzzo

PhD student at Halmstad University. Passionate about meta-learning, few-shot learning, continual learning.

[Dheeraj Varghese @dhevarghese]

Ever wondered whether AI agents could play thousands of video games on a supercomputer? We did. 🎮

Introducing PlaySuite: 5,734 games, 14 open models, and a large-scale benchmark for interactive visual intelligence.

Here's what we built, and what we learned. 🧵👇

[Francesco Locatello @FrancescoLocat8]

And that’s a wrap for the @NeurIPSConf D&B track! 🎉 @konstantina_pll @laroyo @joavanschoren @marco_fumero @AVettoruzzo

[Lora Aroyo @laroyo]

@NeurIPSConf #DBTrack chairs celebration after a year of lots of data work :)) it was a great fun to work with such an awesome team

[Anna Vettoruzzo @AVettoruzzo]

Let’s meet at #ICLR2025 🇸🇬 !

Drop by our poster “Unsupervised Meta-Learning via In-Context Learning” tomorrow at 15.00!

[Anna Vettoruzzo @AVettoruzzo]

Special thanks to my collaborators Lorenzo, @joavanschoren and Marlena.

See you in Singapore!

[Anna Vettoruzzo @AVettoruzzo]

Exciting news 🚀 Our paper "Unsupervised Meta-Learning via In-Context Learning" got accepted at #ICLR2025 !

If you are interested in unsupervised meta-learning and in-context learning have a look at 🐪CAMeLU🐪: https://t.co/0pToMrUJGU

More in 🧵.

[Anna Vettoruzzo @AVettoruzzo]

🔑 We reframe meta-learning as a non-causal sequence modeling problem to achieve cross-domain generalization in several domains.

[Anna Vettoruzzo @AVettoruzzo]

This work was a collaboration with @joavanschoren Mohamed-Rafik Bouguelia, Thorsteinn Rögnvaldsson and it was funded by TAILOR Connectivity Project.

[Anna Vettoruzzo @AVettoruzzo]

I will be at @CoLLAs_Conf in Pisa this week presenting our latest work titled “Learning to Learn without Forgetting using Attention”.

Pass by at our poster on Tuesday 30th!

📄🎥 Paper and video: https://t.co/Mcxj6Kqz8t

#CoLLAs2024

[Anna Vettoruzzo @AVettoruzzo]

TL;DR: A meta-learned optimizer uses attention to learn the relationships between model parameters across a stream of tasks and generate weight updates for the current task while preventing catastrophic forgetting.

[Machine Learning (ML) Papers @Memoirs]

Unsupervised Meta-Learning via In-Context Learning. https://t.co/2qFX10Lr4H

[Anna Vettoruzzo @AVettoruzzo]

Today I had the pleasure to present our recent work ���Meta-Learning from Multimodal Task Distributions Using Multiple Sets of Meta-Parameters” at the International Joint Conference on Neural Networks #IJCNN2023, held in Gold Coast, Australia. Get in touch if you want to know more!

[Anna Vettoruzzo @AVettoruzzo]

Gold Coast here we are! Ready for #IJCNN23. Let’s meet & chat ✨

[Anna Vettoruzzo @AVettoruzzo]

Getting ready for the Workshop on Meta-Learning at #NeurIPS2022 ⏳

[Anna Vettoruzzo @AVettoruzzo]

@thefillm Would love to meetup and chat about meta-learning and continual learning!

[Anna Vettoruzzo @AVettoruzzo]

@andrew_atanov Would love to meetup and chat about meta-learning!

[Anna Vettoruzzo @AVettoruzzo]

Attending the pre-workshop this evening @WiMLworkshop. Let’s get started #NeurIPS2022!

https://t.co/yJRGMEDUyp

[Savannah Thais 👩🏼‍💻 @BasicScienceSav]

It happened! I met some guys in a pub last night, turns out they’re in ML. I was like oh cool what’s your research on? And one of them goes ‘there’s no way you’re in ML, what’s a SVM??’ Guy is a living sexist meme 😳

[Anna Vettoruzzo @AVettoruzzo]

Here I am #NeurIPS22! Let’s meet & chat about meta-learning and distribution shifts and have some fun in New Orleans 🎶