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。

Programming Cognitive Robots

80 点作者 i_don_t_know超过 2 年前

2 条评论

civilian超过 2 年前
I love the idea of cognitive AI. I&#x27;ve dabbled with OpenCog <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;OpenCog" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;OpenCog</a> , and in my youth I read &quot;Artificial General Intelligence (2005)&quot; by Ben Goertzel and I was really convinced.<p>But it seems like... cognitive AI hasn&#x27;t paid off? Big &amp; deep Neural Nets are the type of ML&#x2F;AI that are achieving milestones in learning, gameplay and tasks.<p>If someone has a strong case to make for cognitive programming, I&#x27;d love to hear it. But right now it seems like it&#x27;s a heuristics-based system that&#x27;s destined to lose to NNs.<p>(And before the nitpickers arrive, I totally grant that using heuristics for toy robots makes sense, and is a good way to expose people to programming.)
评论 #32668677 未加载
评论 #32671247 未加载
Animats超过 2 年前
Oh, that feels so retro. That&#x27;s the kind of stuff I was taught going through Stanford in the 1980s, at the beginning of the &quot;AI Winter&quot;.<p>I&#x27;ve actually done a blocks world on a real robot (an IBM RS-1, of all things) using that approach, programming in LISP. Yes, it worked. No, it can&#x27;t deal with an unstructured or dynamic world. Predicate calculus is fine for program proving, but terrible for real world robots.