Search papers, labs, and topics across Lattice.
B-Instruct Qwen3-VL-
3
0
7
0
Even the most advanced MLLMs like GPT-5 and Gemini struggle to spot the "odd one out" in simple visual grids, revealing a surprising weakness in fine-grained visual perception.
VFL label privacy is more fragile than we thought: a novel attack infers labels across multiple scenarios without needing auxiliary data, even bypassing common defenses like gradient noise.
VLMs can get a +39% boost in downstream reasoning by using translator-guided reinforcement learning to improve geometric perception, a far better result than standard supervised fine-tuning.
