正在加载视频...

视频加载失败

What do we really need for social navigation?🤖 Deep RL? No❌ Imitation learning? No❌ LLMs? No❌ 📊Turns out for over 80% of the time, classical methods >> learning-based solutions! In new work, we rethink social navigation. Thread🧶

14,066 次观看 • 2 年前 •via X (Twitter)

5 条评论

Rohan Chandra 的头像
Rohan Chandra2 年前

In typical day-to-day life, you only need to maneuver around other humans in doorways, hallways, intersections about a **fraction** of the time. For the rest, you can "switch off your brain" and just walk straight.

Rohan Chandra 的头像
Rohan Chandra2 年前

But for that tiny fraction where you do need to socially adapt to the situation, classical systems are insufficient. There, you need the representation power of learning methods. so what do you do? just combine the two, and switch between them depending on the situation !

Rohan Chandra 的头像
Rohan Chandra2 年前

We design a new hybrid approach that "targets" the learning only to situations where its needed, defaulting to classical otherwise. It performs better than if using classical or learning alone.

Rohan Chandra 的头像
Rohan Chandra2 年前

📰Preprint: 📽️video: Joint with @_amirraj , Zichao Hu , @KarnanHaresh , Amirreza Payande , Luisa Mao , @PeterStone_TX , @Joydeepb_robots , and @XuesuXiao @texas_robotics @utlarg @amrl_ut

UserInterface 的头像
UserInterface7 年前

The UIX Network is Live! Business & Social Networking 2.0. All the tools you need to get your $ right. By Us For Us. -It's for the people. #getoncode pls rt

相关视频