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。

HMT: Hierarchical Memory Transformer for Long Context Language Processing

87 点作者 jasondavies大约 1 年前

1 comment

cs702大约 1 年前
Code: <a href="https:&#x2F;&#x2F;github.com&#x2F;OswaldHe&#x2F;HMT-pytorch">https:&#x2F;&#x2F;github.com&#x2F;OswaldHe&#x2F;HMT-pytorch</a><p>This looks really interesting. I&#x27;ve added the paper to my reading list and look forward to playing with the code. I&#x27;m curious to see what kinds of improvements we can get by agumenting Transformers and other generative sequence models with this and other mechanisms implementing hierarchical memory.[a]<p>Shouldn&#x27;t the authors cite the work by Jeff Hawkins et al at Numenta? Hawkins has been proposing AI models with hierarchical temporal memory for a long time.[b] I can&#x27;t help but wonder if there is a way, somehow, to incorporate his work and ideas in Transformers and other generative sequence models.<p>We sure live in interesting times!<p>---<p>[a] In the past, I&#x27;ve experimented with mechanisms that add memory to Transformers, but never with <i>hierarchy</i>.<p>[b] <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Hierarchical_temporal_memory" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Hierarchical_temporal_memory</a>
评论 #40392771 未加载
评论 #40395369 未加载