跳到主要內容
AI 武林
影片進階EN2.3 萬 次觀看

Robotics Has Been Stuck for 70 Years — Deepak Pathak, Skild AI

來源 AI Engineer

看影片(在新分頁開啟原站)連到 AI Engineer

摘要

Deepak Pathak 指出機器人學卡在硬體思維,缺乏通用大腦,並展示 Skild AI 的「全身體」模型如何透過模擬、人際影片與遠端運算元據訓練。他說明該模型能跨硬體執行插耳機、煮蛋等任務,並透過部署資料飛輪持續進化,甚至讓斷腿機器人重新學習行走。

Deepak Pathak argues robotics is stuck due to hardware focus and proposes an omni-bodied brain model that scales via simulation, human video, and deployment data flywheels.

摘要、重點與章節標題由語言模型整理,細節(誰說的、數字、先後)可能有誤;要引用請以原始內容為準。

重點

  • 機器人學因過度關注硬體而停滯七十年。
  • Skild AI 提出單一模型處理任何機器人與任務的架構。
  • 透過模擬、人際影片與部署資料的飛輪實現快速進化。

章節

依話題轉折切分,標題由 AI 產生

  1. 00:00AI's progress and the robotics hype
  2. 01:02Robotics has been "almost here" for 70 years
  3. 01:27A 1960s block-copying robot
  4. 02:17Teleoperation in 1957
  5. 03:57Why robotics is stuck: the missing general brain
  6. 04:27Moravec's paradox
  7. 05:42There's no internet of robot data
  8. 07:22Skild's thesis: one brain, any robot, any task
  9. 08:22Teaser: many robots, one brain
  10. 09:31How to scale robot data: a career retrospective
  11. 12:01The three properties of robot data that matter
  12. 13:21Pre-training, post-training and the deployment flywheel
  13. 14:56Why laundry folding is actually easy
  14. 16:17The AirPods task
  15. 17:21Learning from human videos
  16. 18:36The robot's "dream": imagined training data
  17. 18:56Cooking omelets on $4,000 arms
  18. 20:41Truly end-to-end: camera in, motor power out
  19. 21:21Why stairs are harder than backflips
  20. 24:50Deployment: GPU assembly for NVIDIA
  21. 26:00Package delivery to the front door
  22. 26:25Safety: adapting when the robot breaks

提到的工具與公司

  • NVIDIA
  • GPT-3

適合誰看

對機器人學、AI 模型架構或自動化部署感興趣的開發者與研究者。

摘要依據

講者
Deepak Pathak
依據
自動字幕

為什麼排在這裡

人氣
0.83
新鮮
0.96

在主題頁與搜尋結果裡,名次由相關、人氣、新鮮三個分數決定;這一頁沒有搜尋的關鍵字,所以沒有相關分數。排序怎麼算

摘要由 AI 根據原文產生,可能有誤;完整內容請看原站。看影片(在新分頁開啟原站)