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。

Ask HN: Would you like to build a meeting extractor together?

1 点作者 kuro-kuris超过 8 年前
I have recently seen x.ai and I found meeting extraction an extremely interesting technical project for NLP. I would like to extract meetings from texts and return the place and time of a meeting.<p>I need help building a meeting classifier and the initial labeled data set to decide whether documents are meeting invitations or not. To parse the meetings seems a simpler problem but I think there&#x27;s opportunities for some interesting models in this area as well.<p>To this end initially to build a meeting classification model I need some help extracting meeting proposals, I have decided to use the ENRON email data set and the leaked Hillary emails as an initial corpus to build my meeting classifier.<p>After having the meeting invitations I would like to extract and validate the times and places involved.<p>Get in touch if you are interested, email in profile! I think building the meeting corpus is the most time consuming part of the project my current strategy is to think about a list of relevant regex and get a subset of the emails I have based on these and hand mark these to build a labeled data set.

1 comment

sharemywin超过 8 年前
you know it&#x27;s interesting would you consider &quot;Hey, let&#x27;s meet up at Joe&#x27;s for lunch&quot; a meeting?