Aitor González-Marfil retweetledi
while doing ML research i spend a disproportionately large amount of time looking at this specific figure
English
Aitor González-Marfil
17 posts
Aitor González-Marfil
@AAitorG
PhD student | Deep Learning | Computer Vision | Biomedical Images
Katılım Şubat 2022
128 Takip Edilen7 Takipçiler
Aitor González-Marfil retweetledi
🧬 BiaPy was built by an international collaboration between researchers from @upvehu, @DIPCehu, @BiofisikaScienc, @UC3M, @scu_lab, @LabOptBio and others — led by @IgnacioArganda and @ArrateMunoz.
A huge thanks to everyone who contributed! 🌟 (5/7)
English
But there are no restrictions on its use! Feel free to experiment with other types of data and see what works for you. 🚀
English
DINOSim was created to provide a functional tool for environments with limited resources, little data, no labels, or no access to large model training, a common challenge in many biomedical labs.
English
My first preprint is out!
"DINOSim: Zero-Shot Object Detection and Semantic Segmentation on Electron Microscopy Images" 🔬✨
🔗: doi.org/10.1101/2025.0…
English
A few months ago, I had the opportunity to present my project, DINOSim, at #SPAOM2024. It was an incredible experience where I had the opportunity to meet and share ideas with many amazing people.
English
Aitor González-Marfil retweetledi
🧪🔬 Are you at #SPAOM2024 ? Here you have a list of the contributions from my lab (all in Day 2 and 3)👇
1⃣ Tomorrow at 10:15am (Sala Toledo) my PhD student @AAitorG will present his work "Zero-Shot Object Detection with Foundational Models: A Similarity-Based Approach"
English
Aitor González-Marfil retweetledi
#SPAOM2024 attendants don't miss @AAitorG's talk on #ZeroShot #segmentation 👇 and catching him up later!
Ignacio Arganda-Carreras@IgnacioArganda
🧪🔬 Are you at #SPAOM2024 ? Here you have a list of the contributions from my lab (all in Day 2 and 3)👇 1⃣ Tomorrow at 10:15am (Sala Toledo) my PhD student @AAitorG will present his work "Zero-Shot Object Detection with Foundational Models: A Similarity-Based Approach"
English
Aitor González-Marfil retweetledi
🆕 We have updated our preprint in @biorxivpreprint 📰 It explains better the current state of #BiaPy ⛴️ while describing as well its limitations 🤗 Hope you like it!
"BiaPy: Accessible deep learning on bioimages" biorxiv.org/content/10.110…
English
Aitor González-Marfil retweetledi
I remember during my PhD,
I spent many many hours reviewing every single image in my datasets, while I just used an off-the-shelf GAN architecture for my models because I observed that
dataset quality >>>> model arch
Tanishq Mathew Abraham, Ph.D.@iScienceLuvr
English
Aitor González-Marfil retweetledi
DINOv2, the cutting-edge computer vision model trained through self-supervised learning to produce universal features, is now available under the Apache 2.0 license.
Onward with open source AI.
AI at Meta@AIatMeta
Today we’re announcing two new updates in our computer vision work — a new, expanded license for our DINOv2 model and the release of FACET, a comprehensive new benchmark dataset to help evaluate and improve fairness in vision models. More details ➡️ bit.ly/3L35E1U 🧵
English
Aitor González-Marfil retweetledi
I-JEPA: Efficient method for Self-Supervised Learning of image features.
No need for data augmentation, just masking.
Joint embedding predictive architecture, not generative.
And it's open source, of course.
Blog: ai.facebook.com/blog/yann-lecu…
Paper: arxiv.org/abs/2301.08243
Code & models: github.com/facebookresear…
AI at Meta@AIatMeta
Today we're releasing our work on I-JEPA — self-supervised computer vision that learns to understand the world by predicting it. It's the first model based on a component of @ylecun's vision to make AI systems learn and reason like animals and humans. Details ⬇️
English
Aitor González-Marfil retweetledi
GPT-4 "discovered" the same sorting algorithm as AlphaDev by removing "mov S P".
No RL needed. Can I publish this on nature?
here are the prompts I used chat.openai.com/share/95693df4…
(excuse my idiotic typos, but gpt4 doesn't mind anyways)
Jim Fan@DrJimFan
Sorting algorithm underpins all critical softwares. DeepMind's AlphaDev speeds up sorting small sequences (3-5 items) by 70%. Key takeaways: * The main RL algorithm is based on AlphaZero that originally played Go, Chess & Shogi. Same idea applies to searching programs! * Instead of optimizing over C code, they optimize assembly code instead. It's a deliberate choice to go low-level to squeeze out every instruction saving. * The assembly code is then reverse-engineered by hand to C, and open-sourced in LLVM. * Even though the representation network uses transformer, it is NOT a foundation model. The whole pipeline only works on sorting, and has to be re-trained for other tasks like hashing.
English
Aitor González-Marfil retweetledi
Aitor González-Marfil retweetledi
This is a "3D-diffusion" video created using a combination of four different AI models🤯
Welcome to the metaverse! 🌌😎
There's such incredible potential here that I want to explain how I made this, so here's a thread! (1/n)
English
Aitor González-Marfil retweetledi
CLIPasso: Semantically-Aware Object Sketching
abs: arxiv.org/abs/2202.05822
project page: clipasso.github.io/clipasso/
English