AVIS: Autonomous Visual Information Seeking with Large Language Model Agent | Read Paper on Bytez