Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: TLDR This – Auto summarize any article or webpage in a click

tldr.hackeryogi.com

Post reply on HN
Print view View original

81–90 of 117 posts

Re: Show HN: TLDR This – Auto summarize any article or webpage in a click

#81

Post by mgxplyr »

simonebrunozzi wrote: post #79

Good luck! I tried to launch something like this years ago [0] specifically to summarize HN, it's not easy but I think there's value in this type of service. In my case, the summarization was manual - I wanted to see if people liked the service, before building the summarization engine. I also created a chrome extension to be able to see summarize with just one click from HN itself. [0]: https://github.com/simonebrun…

This would be a great type of service for news publishers. Digital producers often have to type out summaries (bullets) in stories for those who don't bother reading full articles. If it could be automated by sending the story copy via an API and getting back summary bullets, it would certainly be valuable.

Re: Show HN: TLDR This – Auto summarize any article or webpage in a click

#82

Post by nothrabannosir »

upheaval7276 wrote: post #60
feanaro wrote: post #58

Earlier quoted context omitted.

Which information? The server is then not reproducing/transmitting/redistributing the content, only indices into the content. I don't see why this would be copyright infringement.

Hacker news is not a good place to get legal advice. Best to ignore anyone offering it. Talk to a lawyer for legal advice.

This comment is so poignant it should be part of the site guidelines.

Re: Show HN: TLDR This – Auto summarize any article or webpage in a click

#83

Post by amiga-workbench »

nudpiedo wrote: post #53

The mankind needs this. Cutting off the BS factory, no more copywritting, flashy headlines with unrelated news boddies, or copy-rewritting, or attention wasters and... the yet to appear ML generated article time wasters which are mash ups from others. If your idea had a company behind I would invest in it. I wish it was a product which would generate a decent enough executive summary, so if I decide to invest more ti…

With the way content on the web is going down the shitter I cant wait to have a program that I run on a remote server, one that puppets around a headless browser, executes all the usual javascript cancer and reinterprets the final pages into sane, calm summarised document for local rendering.

Re: Show HN: TLDR This – Auto summarize any article or webpage in a click

#84

Post by p1esk »

steve19 wrote: post #22
hacker089 wrote: post #18

Sorry, i've some experience in this field: - This is text-extraction, NOT text-generation - TextRank algorithm is so far fine, but it does not write a "summary", instead it ranks the components of a text according to some "metrics" (simply spoken) - Using this approach will still make you attackable by copyright claims from copyright owners - Which stuff is summarized ("put to the final output") is not always clear t…

I'm interested in learning about true generation algorithms. Can you point me in the right direction?

Google “gpt-2”.

Re: Show HN: TLDR This – Auto summarize any article or webpage in a click

#85

Post by Der_Einzige »

bbmario wrote: post #67

Any open-source libraries that do this, or anything similar to this?

I wrote a tool for this called CX_DB8. It's better than any other "TextRank" based extractive summarizer in that it utilizes the latest in word embeddings, and supports queryability. Furthermore, it is the only word-level extractive summarizer to support all of this. Mostly aimed at the competitive debate community but anyone can use it.

https://github.com/Hellisotherpeople/CX_DB8

Hopefully you'll see my paper about this in EMNLP 2019 System Demos...

Re: Show HN: TLDR This – Auto summarize any article or webpage in a click

#86

Post by Der_Einzige »

Shameless self plug, but for anyone interested in the SOTA of extractive text summarization, I wrote a tool for this which allows for folks to use any type of word embeddings to run a modified queryable text rank on. This leads to SOTA text extraction.

https://github.com/Hellisotherpeople/CX_DB8

CX_DB8 also supports word and sentence level extraction. I plan on adding paragraph extraction in a future update.

Re: Show HN: TLDR This – Auto summarize any article or webpage in a click

#88

Post by radhakrsna »

blululu wrote: post #66

Tried to use this service on the HN comments discussing it. It could not generate a summary of HN comments, now I need to read people's opinions the hard way.

Hi, It is mainly designed for articles so it wouldn't give any useful info when used on a page like HN. Thank you very much for taking the time to use the app and provide your feedback.

Re: Show HN: TLDR This – Auto summarize any article or webpage in a click

#90

Post by radhakrsna »

Der_Einzige wrote: post #85
bbmario wrote: post #67

Any open-source libraries that do this, or anything similar to this?

I wrote a tool for this called CX_DB8. It's better than any other "TextRank" based extractive summarizer in that it utilizes the latest in word embeddings, and supports queryability. Furthermore, it is the only word-level extractive summarizer to support all of this. Mostly aimed at the competitive debate community but anyone can use it. https://github.com/Hellisotherpeople/CX_DB8 Hopefully you'll see my paper about…

Ah sounds good. I might give it a try later on. Thanks for sharing.
Post reply on HN