Jhony H. Giraldo
banner
jhonyhgiraldo.bsky.social
Jhony H. Giraldo
@jhonyhgiraldo.bsky.social
Assistant professor at Télécom Paris, Institut Polytechnique de Paris 🇨🇵🇨🇴. #GeometricDeepLearning #ComputerVision https://jhonygiraldo.github.io/
We're also presenting our work "Subgraph Gaussian Embedding Contrast for Self-Supervised Graph Representation Learning", accepted in the Research Track.

📄 Paper: arxiv.org/abs/2505.23529

📌 Join us:

-Poster: Tuesday, 7–9 PM.
-Oral: Wednesday, 4:12–4:30 PM (Representation Learning II session).
Subgraph Gaussian Embedding Contrast for Self-Supervised Graph Representation Learning
Graph Representation Learning (GRL) is a fundamental task in machine learning, aiming to encode high-dimensional graph-structured data into low-dimensional vectors. Self-Supervised Learning (SSL) meth...
arxiv.org
September 14, 2025 at 8:47 PM
I have the feeling some journals are replacing "major revisions" with "rejection with possible resubmission" to artificially reduce acceptance rate. I'd resubmit with a letter explaining why the suggestion is not valid or run the additional baseline (not sure though if it's an additional baseline).
September 3, 2025 at 4:24 PM
Reposted by Jhony H. Giraldo
Let's push for the obvious solution: Dear @neuripsconf.bsky.social ! Allow authors to present accepted papers at EurIPS instead of NeurIPS rather than just additionally. Likely, at least 500 papers would move to Copenhagen, problem solved.
August 28, 2025 at 7:19 PM
Yes, this is unclear @euripsconf.bsky.social
July 17, 2025 at 6:30 AM
We could also put some effort in the conferences happening here in Europe, e.g. ECML, BMVC, GCPR, etc. It's also up to us to some extent.
March 20, 2025 at 8:04 PM
💡 I won’t be at AAAI or WACV, but my PhD students will! If you're attending, don’t miss their talks & posters—reach out & connect!

With support from @telecomparis.bsky.social, @ipparis.bsky.social, Univ. of Turin, @udea.bsky.social, EIDOS Lab, Hi! PARIS Center & ANR.
February 24, 2025 at 9:39 PM
CATALOG leverages contrastive learning + multi-modal foundation models to enhance wildlife species recognition in challenging real-world conditions. 🐆📸 Catch it at @wacvconference.bsky.social!
February 24, 2025 at 9:39 PM
🔹 CATALOG: A Camera Trap Language-Guided Contrastive Learning Model
📍 #WACV2025 – Tucson
📅 Oral: Mar 2, 4:00 PM - 5:00 PM (AZ Ballroom Salon 6)
📅 Poster: Mar 2, 11:15 AM - 1:00 PM (Tucson Ballroom, Paper ID 136)
📖 arxiv.org/abs/2412.10624 | 💻 github.com/Julian075/CA...
CATALOG: A Camera Trap Language-guided Contrastive Learning Model
Foundation Models (FMs) have been successful in various computer vision tasks like image classification, object detection and image segmentation. However, these tasks remain challenging when these mod...
arxiv.org
February 24, 2025 at 9:39 PM
WiGNet solves scalability issues in vision GNNs with a windowed processing approach, reducing computational cost while keeping top-tier performance. Don't miss this at @wacvconference.bsky.social!
February 24, 2025 at 9:39 PM
🔹 WiGNet: Windowed Vision Graph Neural Network
📍 #WACV2025 – Tucson
📅 Oral: Mar 2, 10:15 AM - 11:15 AM (AZ Ballroom Salon 8–12)
📅 Poster: Mar 3, 11:15 AM - 1:00 PM (Tucson Ballroom, Paper ID 43)
👨‍🏫 Presented by: Gabriele Spadaro
📖 arxiv.org/abs/2410.00807 | 💻 github.com/EIDOSLAB/WiG...
WiGNet: Windowed Vision Graph Neural Network
In recent years, Graph Neural Networks (GNNs) have demonstrated strong adaptability to various real-world challenges, with architectures such as Vision GNN (ViG) achieving state-of-the-art performance...
arxiv.org
February 24, 2025 at 9:39 PM
HYGENE is the first diffusion-based method for hypergraph generation, enabling realistic modeling of complex, higher-order relationships. 🔥 If you're at #AAAI2025, don't miss it!
February 24, 2025 at 9:39 PM
🔹 HYGENE: A Diffusion-Based Hypergraph Generation Method
📍 #AAAI2025 – Philadelphia
📅 Feb 28, 12:30 PM - 2:30 PM (Paper ID 270)
👨‍🏫 Presented by: Dorian Gailhard
📖 arxiv.org/abs/2408.16457 | 💻 github.com/DorianGailha... | 📝 doriangailhard.github.io/blog/2024/HY...
HYGENE: A Diffusion-based Hypergraph Generation Method
Hypergraphs are powerful mathematical structures that can model complex, high-order relationships in various domains, including social networks, bioinformatics, and recommender systems. However, gener...
arxiv.org
February 24, 2025 at 9:39 PM