Pavel Šuma

PhD student @ Visual Recognition Group, CTU in Prague 🇨🇿
Computer Vision and chill

[Bill Psomas @bill_psomas]

🇧🇷 Presenting our ICLR 2026 paper “Efficient Probing” (EP) today!

❓What if linear probing is asking the wrong question?

🥳 EP is a lightweight attention probing method that better evaluates local, patch-level representations from models like MIM.

📍P4-#3713, 15:15–17:45

[Pavel Šuma @SumaPavel]

🚨 Efficient Local Visual Similarity (ELViS) @ #ICLR 2026 🇧🇷

ELViS is a fast, lightweight, and interpretable module for estimating image-to-image similarity that generalizes well to many image domains.

Paper: https://t.co/OEEDqPHwni
Come see poster today @ P4-#3715

[Giorgos Kordopatis-Zilos @g_kordo]

ILIAS is a large-scale dataset for evaluation on Instance-Level Image retrieval At Scale. It is designed to support research in image-to-image and text-to-image retrieval for particular objects and serves as a benchmark for evaluating foundation models and retrieval techniques