가입 후 초대 링크를 공유하면 동영상 재생 및 초대 보상을 받을 수 있습니다.

Biwei Huang
@huang_biwei
Founder @AetherLab_AI Assistant Professor @HDSIUCSD @UCSanDiego Causal World Model, Causality-driven Agentic System for the next AI paradigm
가입 September 2020
255 팔로잉 중    5.8K 팬
The hard part often comes one step earlier: how do we recover the relevant causal variables in the first place? In robotics, these variables are often not given directly. They have to be learned from inputs such as pixels, trajectories, and raw sensor streams before we can model how actions change the system.
더 보기
What makes a world model causal? As a robotic gripper closes around an object, the contact conditions change. Friction, grasp point, speed and angle can all affect the outcome. Our framework treats causality as an explicit system of causal variables, causal structure and causal dynamics in latent space. Which condition would you vary first? #CausalWorldModels#
더 보기