Wei He

Dense pixel-level segmentation can be performed directly inside an autoregressive token space without external decoders by serializing binary change masks into grammar-constrained quadtrees.

Key contribution not extracted.

MLLMs excel at reasoning over diagrams but falter in parsing and editing them, revealing a significant gap that needs addressing.

CogVis redefines change detection by decoupling temporal perception from semantic categorization, achieving unprecedented efficiency and accuracy across multiple benchmarks.

DIVE achieves an impressive 88.9% reduction in visual tokens without sacrificing performance, redefining efficiency in vision-language models.

Adding depthwise convolutions to Transformers can boost accuracy on downstream tasks while barely increasing model size.