1 paper
Haichen He, Jiayi Zhou, Sifeng Shang +3
Real-world long video understanding requires models to perform continuous tracking, information integration and memory retention over massive temporal spans within extreme video du…