1 paper
Songtao Jiang, Sibo Song, Chenyi Zhou +12
The transition from image to video understanding requires vision-language models (VLMs) to shift from recognizing static patterns to reasoning over temporal dynamics such as motion…