TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Building AGI Using Language Models

26 点作者 leogao超过 4 年前

5 条评论

hprotagonist超过 4 年前
I remain to be convinced that encoding statistical information about syntactical manipulation alone will somehow magically convert to semantic knowledge and agency if you just try really hard and do it a lot.
评论 #24194990 未加载
评论 #24203941 未加载
评论 #24194827 未加载
评论 #24194817 未加载
评论 #24195009 未加载
ilaksh超过 4 年前
To me GPT-3 excitement is equivalent to when people get hyped about &quot;defeating aging&quot; after seeing some resveratrol trial or something.<p>Language is only part of it. And you can&#x27;t get complete understanding without integrating spatial information. Take a look at Josh Tenenbaum&#x27;s work for explanation of why.
msamogh1超过 4 年前
It talks about how you can go from a language model that simply generates text to an agent that is capable of performing actions in the real world<p>Essentially, the missing pieces in the picture come down to input and output modules. &quot;How do you formulate any given problem into a form that a language model can answer?&quot;.
评论 #24194972 未加载
bionhoward超过 4 年前
If the agent state is symbolic, that’s cool and interesting, but isn’t reality sub-symbolic?
mrfusion超过 4 年前
I don’t completely follow this. Can anyone explain?
评论 #24194907 未加载