Menu
Sign In Search Podcasts Charts People & Topics Add Podcast API Pricing
Podcast Image

AI可可AI生活

[人人能懂] 从学会走路,到打通游戏,再到过目不忘

30 Nov 2025

Description

你有没有想过,AI的“开窍”瞬间,背后藏着什么样的秘密?这一期,我们将一口气解锁五篇最新论文,探讨AI智能的进化法则。我们会发现,为什么让AI的大脑长得“更高”而不是“更胖”,能让它学会体操一样的神操作;为什么“学得快”的AI画家比“学得久”的更有创造力;以及AI是如何通过自我复盘和深度研究,从一个模仿者,成长为能打通复杂游戏的通才。00:00:33 人工智能的成长秘密:与其更胖,不如更高00:04:46 AI画画,为什么“学得快”的比“学得久”的更聪明?00:09:32 不止是模仿:机器人如何学会自我迭代00:14:42 AI上岗:我们是怎么教会它打通一款游戏的?00:20:44 你的大脑不是硬盘,而是搜索引擎本期介绍的几篇论文:[LG] 1000 Layer Networks for Self-Supervised RL: Scaling Depth Can Enable New Goal-Reaching Capabilities[Princeton University]https://openreview.net/forum?id=s0JVsx3bx1---[LG] Why Diffusion Models Don't Memorize: The Role of Implicit Dynamical Regularization in Training[Université PSL]https://arxiv.org/abs/2505.17638---[LG] π∗0.6: a VLA That Learns From Experience[Physical Intelligence]https://arxiv.org/abs/2511.14759---[AI] Lumine: An Open Recipe for Building Generalist Agents in 3D Open Worlds[ByteDance Seed]https://arxiv.org/abs/2511.08892---[CL] General Agentic Memory Via Deep Research[Beijing Academy of Artificial Intelligence]https://arxiv.org/abs/2511.18423

Audio
Featured in this Episode

No persons identified in this episode.

Transcription

This episode hasn't been transcribed yet

Help us prioritize this episode for transcription by upvoting it.

0 upvotes
🗳️ Sign in to Upvote

Popular episodes get transcribed faster

Comments

There are no comments yet.

Please log in to write the first comment.