Adrian Arnaiz

A Sociotechnical Approach to Trustworthy AI: from Algorithms to Regulation This thesis presents a sociotechnical framework for implementing Trustworthy Artificial Intelligence (TAI), integrating technical, human, and regulatory aspects. It emphasizes the importance of aligni...

The thesis document is available here 📖 ellisalicante.org/publications...

I've defended my PhD 🎓🙂

Huge thanks to @nuriaoliver.com for her guidance over these years and to the committee @ccattuto.bsky.social , Isabel Valera, and G. González Serrano for their time, insights, and an inspiring discussion 🙏

Excited for what's coming next!

EU Tech Policy Brief: September 2025 Welcome back to the Centre for Democracy & Technology Europe‘s Tech Policy Brief! This edition covers the most pressing technology and internet policy issues under debate in Europe and gives CDT’s per...

📢 The latest EU Tech Policy Brief is here!

In this edition:

⚖️ Joint letter to EU Commission & MSs on the #AIAct ’s implementation

👁️ Danish Presidency’s proposal on EU #CSAR raising concerns on #encryption

💬 Our response to the Gender Equality Strategy’s consultation

👇 More about EU tech policy:

Ladies and gentlemen, the generative AI industry:

No to simplification of EU digital laws

Weakening the AI Act would undermine accountability, the protection of society and workers and bring legal uncertainty

I have co-signed the letter led by @cdteu.bsky.social 👇🏼

The Code of Practice is out. I co-wrote the Safety & Security Chapter, which is an implementation tool to help frontier AI companies comply with the EU AI Act in a lean but effective way. I am proud of the result!
1/3

A huge milestone was achieved today in the EU towards our common understanding of risks in the online world. If you are a researcher interested in studying societal risks related to very large online platforms and search engines, this is for you!
#DSA
digital-strategy.ec.europa.eu/en/news/comm...

Needing a permanent doi was a forcing function to revisit this article on Transformers and Graph Neural Networks: arxiv.org/abs/2506.22084

This is (very) old news, so what's new?

Had a great time presenting our work at @icwsm.bsky.social on measuring and mitigating structural social capital disparities in social networks 🕸️

Really appreciated the thoughtful feedback and the chance to connect with such an amazing community!

📄 Read the paper: ojs.aaai.org/index.php/IC...

This paper reflects my and @federico-errica.bsky.social's thoughts on some current GNNs issues (like OSM/OSQ) after spending some time working on them. We also tried to capture concerns that we've seen echoed by parts of the community.

P.S. I thought I had all these issues figured out too 😅

Oversmoothing, "Oversquashing", Heterophily, Long-Range, and more: Demystifying Common Beliefs in Graph Machine Learning After a renaissance phase in which researchers revisited the message-passing paradigm through the lens of deep learning, the graph machine learning community shifted its attention towards a deeper and...

@neuripsconf.bsky.social what is going on...

We were desk-rejected from NeurIPS for dubious reasons, but that is not going to stop us 🌶️

arxiv.org/abs/2505.15547

Adrian Arnaiz-Rodriguez, Federico Errica: Oversmoothing, "Oversquashing", Heterophily, Long-Range, and more: Demystifying Common Beliefs in Graph Machine Learning https://arxiv.org/abs/2505.15547 https://arxiv.org/pdf/2505.15547 https://arxiv.org/html/2505.15547

After a break in April we welcome all of you to the next session of GLOW🌟 next week!

Join and interact with our speakers Christian Koke (scale in GNNs) and Yonatan Sverdlov (sparse geometric MPNNs).

🗓️ May 28th, 5pm CEST on Zoom.

🌐 Details & sign-up: sites.google.com/view/graph-learning-on-weds.

I've recently put together a "Fairness FAQ": tinyurl.com/fairness-faq. If you work in non-fairness ML and you've heard about fairness, perhaps you've wondered things like what the best definitions of fairness are, and whether we can train algorithms that optimize for it.

GitHub - adrianjav/causal-flows: CausalFlows: A library for Causal Normalizing Flows in Pytorch CausalFlows: A library for Causal Normalizing Flows in Pytorch - adrianjav/causal-flows

Have you ever been curious to try Causal Normalizing Flows for your project but found them intimidating? Say no more 😜

I just released a small library to easily implement and use causal-flows:

github.com/adrianjav/ca...

If you are interested in contributing to the Graph Learning community, joining the LoG organising team is a great way to do it! 🧠🕸️

🚨 We’re recruiting organizers for LoG 2025!

Join us in shaping the next Learning on Graphs conference at UCLA this fall.

Passionate about graph ML & building community? Apply by Apr 11 (AOE) 🌍

📋 Form: docs.google.com/forms/d/e/1F...

🏠LoG website: logconference.org

Want to know about the current understanding of the generalization abilities of GNNs? Please have a look at our survey paper arxiv.org/abs/2503.15650.

Joint work with Antonis Vasileiou, Stefanie Jegelka, and Ron Levie.

Our work, "Privacy and Accuracy Implications of Model Complexity and Integration in Heterogeneous Federated Learning" (w\ Miguel Angel Lozano, Novi Quadrianto, and @nuriaoliver.com ) is now published at @ieeeaccess.bsky.social: ieeexplore.ieee.org/document/109...

We welcome continued feedback and discussion as we head into the final stages of drafting.

LOGML 2025 London Geometry and Machine Learning Summer School, July 7-11 2025

🌟Applications open- LOGML 2025🌟

👥Mentor-led projects, expert talks, tutorials, socials, and a networking night
✍️Application form: logml.ai
🔬Projects: www.logml.ai/projects.html
📅Apply by 6th April 2025
✉️Questions? logml.committee@gmail.com

#MachineLearning #SummerSchool #LOGML #Geometry

Stylized map of Detroit (MI) showing the highway network, and the network of social connections between urban residents. The connections intersecting highways are sparser than elsewehere. Image credit Karo Berghuber (Insta: @kariot.lines)

"Urban Highways Are Barriers to Social Ties" out on PNAS!
The 1st large-scale measure of how highways weaken social connections between the communities they separate. This barrier effect is strong in the 50 largest US cities--especially for low-income Black communities.
www.pnas.org/doi/10.1073/...

12 mandatory free-text input fields in the review form for ICML? Really?

I am hiring for a Ph.D. student to work in the areas of social network analysis, algorithms and fair machine learning.

Please apply and join our highly motivated team.

For more information please see the call: neumannstefan.com/hiring/

An interesting future resource to keep an eye on will be the resulting report of Working Group 4 on the development of a General-Purpose AI Code of Practice, which will focus on mitigating the risk of GPAI models through AI governance

digital-strategy.ec.europa.eu/en/news/meet...

really looking forward to it!

GLOW is back this Wednesday, Feb 19th, 5pm CET 🌟

Instead of talks, we’re hosting a casual discussion on where graph learning is headed and where early-career researchers can make the most impact. Join our open conversation with junior and senior graph researchers!

sites.google.com/view/graph-l...

And so we set out to understand _feedforward_ graphs (i.e. graphs w/o back edges) ⏩

Turns out these graphs are rather understudied for how often they are used in practice! 😮

We work towards a framework to analyse them, which I hope both my GNN and my LLM friends will enjoy 🚀

Statement from Dario Amodei on the Paris AI Action Summit A call for greater focus and urgency

While the Summit was the scene of important discussions, notably about innovations in health & environment, these promises will only materialize if we address with realism the urgent question of the risks associated with the rapid development of frontier models.
www.anthropic.com/news/paris-a...
1/2

Today, we are publishing the first-ever International AI Safety Report, backed by 30 countries and the OECD, UN, and EU.

It summarises the state of the science on AI capabilities and risks, and how to mitigate those risks. 🧵

Full Report: assets.publishing.service.gov.uk/media/679a0c...

1/21