AIR's first summer camp SUMMER AIR successfully ended. During the summer camp which lasted for two months, students experienced the frontier technology of artificial intelligence and participated in many academic sharing sessions online and offline. The research topics covered different directions involving artificial intelligence, robotics and machine learning, such as graphics, computer vision, recommendation algorithm and exponential lift-off.

The summer camp was organized around four major themes: artificial intelligence, robotics, machine learning and big data. The students focused on one of these areas or cross-cutting areas to carry out intensive research and practice, and the topics carried out included: implicit complementation study of autonomous driving scenes, 3D reconstruction in single color images, self-supervised depth estimation in road scenes, navigation study of home service robots, accurate pouring of liquids based on robotic arms, and so on.

The champion of Summer AIR camp was won by Beiwen Tian, who, together with AIR intern Liyi Luo, introduced the self-supervised learning method Barlow Twins, which attracted a lot of attention in the first half of the year, into the field of 3D learning, improving it into a simple but effective method ——ViewpointBottleneck.

Read More: https://mp.weixin.qq.com/s/OwpfShqPN2A6ak2R3FU2SA