
AI Researcher at MediaTek Research | PhD from @UniPadova, prev. intern at Meta AI, Samsung AI | Graph Representation Learning | Graph Neural Networks.
Joined September 2019
- Tweets90
- Following333
- Followers182
- Likes180
Davide Buffelli retweeted
I will be presenting our paper on deep equilibrium algorithmic reasoning (poster sess 3)
Feel free to swing by, say hi, have a coffee break with me in between sessions or suggest me good locations for a photo
#NAR #NeurIPS2024 #NeurIPS
Davide Buffelli retweeted
In < 1h, we will be taking the LoG stage for the new and improved version of our NAR tutorial 🔢!
Hope you can join us -- it is open for all, publicly streamed on YouTube, and will feature a fun discussion of great reasoning research over both graphs 🕸️ and language 💬!
🌟 Day 4 of LoG 2024 – The Grand Finale!
🎙️ Keynote: @AldenHung ( @IsomorphicLabs )
💻 Tutorial: Neural Algorithmic Reasoning II: From Graphs to Language
🏆 Oral Presentations: TMLR Track
Join us for an exciting final day of insights and innovation! 🚀
Davide Buffelli retweeted
The tutorial will feature many works from our group over the past 2 years, as well as many fantastic papers from folks we admire: @KeyuluXu @jingling_li @tanghao95 @PeterWBattaglia @Montgomerybohde @ShuiwangJi @lrjconan @LProkhorenkova @DobrikG @DBuffelli @sonjj74 @mmbronstein...
Davide Buffelli retweeted
📢 Exciting News! Our paper, “Bayesian Computation Meets Topology,” has just been published in TMLR! 🎉
👉openreview.net/forum?id=0h1D… 👈
Here’s a deep dive into how #topology and #bayes(ian) computation come together to enhance parameter inference:
📌 Why Topology?
Topology provides powerful descriptors like persistent homology that capture the “shape” of data across scales. Our approach leverages these topological features in Bayesian inference—crucial for scenarios where data is scarce or structurally complex.
📎 The Gap We Address
Topological methods are robust for capturing the overall shape of data but don’t integrate into Bayesian frameworks in a straight-forward fashion. We’ve developed a method that fills this gap, incorporating topology-based loss functions for Bayesian parameter estimation.
🛠️ Core Methodology
Our framework uses topological loss functions based on persistent homology for inference, enabling us to construct a comparison-based posterior. This allows for uncertainty quantification without needing explicit likelihoods—essential for chaotic systems where the likelihood is analytically intractable.
🚀 Applications & Experiments
We validate our approach on models with inherent complexity, such as the Vicsek swarm model and Lattice Boltzmann simulations. In these cases, topology-driven Bayesian inference produced significantly more accurate parameter estimates than standard geometry-based ones.
🧩 Implications
This work brings Bayesian computation and topology closer together, paving the way for robust, simulation-based inference. Applications extend to fields like biology and physics, where data often exhibit complex, multi-scale structures that benefit from this topologically-informed approach.
👀 Check It Out!
📜 openreview.net/pdf?id=0h1DtR…
💻 github.com/aidos-lab/TABAC
Joint work with @JRohrscheidt and @SeBayesian!
Davide Buffelli retweeted
#neurips2024 will be very DEAR to me this year!
Deep Equilibrium Algorithmic Reasoning has also been accepted at NeurIPS after its selection for an oral presentation at MAR@CVPR 2024. Special thanks to my collaborators @DBuffelli @pl219_Cambridge and, my lucky charm, @sonjj74!
Davide Buffelli retweeted
📢 I'm officially on the job market📢
If you know me and like what I do, get in touch or help with a retweet, please!
If you do not know me, I ❤️ RL and graphs, especially for practical applications)
My CV: notion.so/whoami-495d6335cf9…
Davide Buffelli retweeted
2020's most cited paper across all scientific disciplines was (still) 2015's ResNet.
Paper: https://nitter.cf/t.co/mGGX1OpseY (v/MSFTResearch)
More: https://nitter.cf/t.co/e3Va7P6Ea8 (@TDataScience)
#DeepLearning #BestOf2020
#WednesdayWisdom
Davide Buffelli retweeted
Here's a thread on lesser-known tools and packages that I could not live without, starting with Python.
(suggestions are very welcome!)
einops:
- github.com/arogozhnikov/eino…
- Easily-interpretable reshapes + tiling + aggregations for numpy/torch/tf/etc
1/n
Davide Buffelli retweeted
🎉 Papers with Code partners with arXiv! Code links are now shown on arXiv articles, and authors can submit code through arXiv. Read more: medium.com/paperswithcode/pa…
Davide Buffelli retweeted
Are you taking full advantage of Python 3?
Are you sure?
Here are 10 Python 3 features that will change the way you are writing code today.
🧵👇
Davide Buffelli retweeted
NeurIPS 💯🎉!
“all come all served”
Registration is $25 for students and $100 for non-students.
Tutorials, keynotes and oral pre-recorded presentations accessible without a registration. Video recordings of the poster presentations will be released after the conference.
Read about changes to conference registration this year, which opens just days from now!
medium.com/@NeurIPSConf/neur…
Davide Buffelli retweeted
I'd love @NeurIPSConf @icmlconf to adopt @openreviewnet bc it'd provide a great discussion forum for authors, reviewers and the AI *community*.
I'd feel mentally better if I knew my answers could be seen/discussed by all researchers interested in my project.
This Post is from an account that no longer exists. Learn more
Davide Buffelli retweeted
We are releasing, #OpenSelfSup, an open-source library for self-supervised learning:
github.com/open-mmlab/OpenSe…
- *All methods in one repository*, supporting PIRL, MoCo, SimCLR, etc.
- *All benchmarks in one repository*.
- *Efficiency*, supporting multi-GPU distributed training.
Davide Buffelli retweeted
Happy to announce that our survey on Representation Learning for Dynamic Graphs is published at JMLR (the Journal of Machine Learning Research). Thanks to all my co-authors/colleagues @BorealisAI.
Link to the paper: jmlr.org/papers/v21/19-447.h…
Davide Buffelli retweeted
TorchServe and [TorchElastic for Kubernetes], new libraries for serving and training models at scale. Learn more: https://nitter.cf/t.co/j6tSOaG6yU
Davide Buffelli retweeted
Facebook AI and AWS have partnered to release libraries that target high-performance PyTorch model deployment and large-scale model training. These enable the community to push the state of the art as model architectures increase in size and complexity. ai.facebook.com/blog/faceboo…
Davide Buffelli retweeted
Up on arXiv: my manuscript about the theory behind Group CNNs, covering recent papers with a consistent notation, including missing proofs and alternative viewpoints. Check it out! arxiv.org/abs/2004.05154
Davide Buffelli retweeted
"You know, people think mathematics is complicated. Mathematics is the simple bit. It's the stuff we can understand. It's cats that are complicated". RIP John Conway
Davide Buffelli retweeted
Our ***free deep learning book*** (d2l.ai) now has a ***modern NLP chapter draft***, complete w. the omnipresent BERT & friends. All drafted via interactive Jupyter notebooks.
We have re-organized Chapter: NLP pretraining (d2l.ai/chapter_natural-langu…) and Chapter: NLP applications (d2l.ai/chapter_natural-langu…), and added sections of BERT (model, data, pretraining, fine-tuning, application) and natural language inference (data, model).
Davide Buffelli retweeted
Been looking to learn about flow models?
Full package released:
- lecture: youtu.be/JBb5sSC0JoY
- slides: https://nitter.cf/t.co/1khtfMbLEV
- lecture colab: https://nitter.cf/t.co/VEhG5q8z0L
- homework: https://nitter.cf/t.co/dxXQE60CiU
w/Alex Li, Wilson Yan, @Aravind7694, @peterxichen, @hojonathanho



































