iiinova — where your digital self keeps evolving

Mert Bulent Sariyildiz

@mbsariyildiz.bsky.social

185 followers 134 following 7 posts

Research scientist at Naver Labs Europe. mbsariyildiz.github.io

Mert Bulent Sariyildiz @mbsariyildiz.bsky.social · 01/09/2026

🧐 Can you distill knowledge from 4 vision teachers into one student using ZERO real images? 🥳 Turns out: YES, and surprisingly well. 📣 Our ECCV 2026 paper "IDeaL" closes most of the gap with real-image distillation, using optimized structured noise.

1279

Reposted by Mert Bulent Sariyildiz

Pau de Jorge @pdejorge.bsky.social · 27/07/2026

1/6 Excited to share that our paper on model merging was accepted at ECCV 2026! 🎉 We introduce an efficient, decoder-free proxy that makes model selection faster, simpler and practical across vision tasks. 📄 arxiv.org/abs/2604.12935 🌐 europe.naverlabs.com/task-alignment 🧵👇

1146

Reposted by Mert Bulent Sariyildiz

Christian Wolf @chriswolfvision.bsky.social · 23/06/2026

For your Embodied AI task you want a recurrent model with constant complexity per step, but you don't want to lose the power of transformers (which store the full obs history and attend to it)? Do not despair, we have your back. We distill transformers into recurrent transformers 1/8

23410

Reposted by Mert Bulent Sariyildiz

Philippe Weinzaepfel @weinzaepfelp.bsky.social · 15/06/2025

Wanna the outstanding performance of MASt3R while using a ViT-B or ViT-S encoder instead of its ViT-L one? Don't miss how we build DUNE, a single encoder for diverse 2D & 3D tasks, at this afternoon #CVPR2025 poster session (poster #376). paper: arxiv.org/abs/2503.14405 code: github.com/naver/dune

0176

Reposted by Mert Bulent Sariyildiz

Yannis Kalantidis @skamalas.bsky.social · 09/06/2025

🧵 Two new papers at #CVPR2025 on generalization in visual representations — covering both universal encoders for 2D and 3D tasks, and open-vocabulary semantic segmentation Let's dive in! 👇

1233