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

Can you read this cursive handwriting? The National Archives wants your help

smithsonianmag.com

Post reply on HN
Print view View original

151–160 of 267 posts

Re: Can you read this cursive handwriting? The National Archives wants your help

#151

Post by dylan604 »

Daneel_ wrote: post #143
Unearned5161 wrote: post #78

Ok I did one letter, from a woman in 1814 writing to James Monroe (then Secretary of State) asking for a passport to go to Scotland to get her late brother's property. What a trip! So enjoyable to get into the flow once you've "synchronized" with the persons handwriting. Furthermore, due to the fact that you're reading and re-writing word for word of whatever you're transcribing, the stories you end up reading have t…

I wish this technique worked for me. I can transcribe something verbatim and then have absolutely no idea what I've written - I have to go back and read it to actually parse the text.

That’s not uncommon. I was the same way back when I took an actual typing class. The part of my brain used for storage/recall just seems to go to sleep when doing the whole transcription stage. Maybe it was a mental thing realizing it was just a task and no actual interest in the content other than accomplishing a task vs doing it something I had a vested interest???

Re: Can you read this cursive handwriting? The National Archives wants your help

#152

Post by patrick451 »

dr_dshiv wrote: post #128
patrick451 wrote: post #121

Earlier quoted context omitted.

The incident with the lawyers just highlighted the fundamental problem with LLMs and AI in general. They can't be trusted for anything serious. Worse, they give the apppearence of being correct, which leads humans "checkers" into complacency. Total dumpster fire.

Instead of thinking about this as an all-or-nothing outcome, consider how this might work if they were made accessible with LLMs, and then you used randomized spot checks with experts to create a clear and public error rate. Then, when people see mistakes they can fix them. I’m trying to do this for old Latin books at the Embassy of the Free Mind in Amsterdam. So many of the books have never been digitized, let alone…

These kinds of ideas just sound to me like "Suppose you had to use broken technology X. How do you make work?"

Re: Can you read this cursive handwriting? The National Archives wants your help

#153

Post by jncfhnb »

CamperBob2 wrote: post #92
tedunangst wrote: post #63

Earlier quoted context omitted.

Can someone please post a sample of one of these images that can only be read by a human for us naive OCR believers to see?

To be fair there was a similar discussion a few days ago in which an SME remained unconvinced: https://news.ycombinator.com/item?id=42566391 I don't necessarily agree with her conclusion because she wasn't participating directly in the thread and wasn't completely responsive to some of the points raised, but still, it appears that there are a few instances of difficult-to-read handwriting where OCR is still coming in…

That’s comprehension of English not reading characters

Re: Can you read this cursive handwriting? The National Archives wants your help

#154

Post by awithrow »

kopirgan wrote: post #126

Is that true?! US kids don't learn cursive? How do they write?!

I'm in the US and learned it in school. I just never really needed to use it consistently. Assignments and papers that were still handwritten could be done either way. Cursive never felt noticeably faster for me to write. I'm sure it would have had I been forced to do it. By the time I was in high school (1999), i remember typing most long form assignments. Now the only time I ever read cursive is on letters from my mom and her cursive is not particularly neat or clean.

Re: Can you read this cursive handwriting? The National Archives wants your help

#155

Post by wriggler »

seletskiy wrote: post #109

To tptacek and other guys who seem to have unwavering trust in OCRs/LLMs, as well as to opposite party who think that technology is not there yet — you are all partially right, but somehow fail to hear each other while also spending time on baseless arguing instead of factual examples and attempts to find common truth. Can it be used to greatly simplify efforts by getting through boilerplate? — Yes. Should the result…

From https://www.handwritingocr.com - seemed to be more accurate, mostly getting the New Mexican and possibly other parts:

---

and I don't know whether it can be reset for a date in December or not. Cornell seemed anxious that it should not come off too close to Christmas, and of course New Mexican practice would support him. I will take this up with the Judge and with Hanna the moment I can get rid of the brief. Meanwhile I would like to know whether there is anything else in which I can be useful to you, since it behooves me to be diligent in view of uncomfortable relations with the present management.

Are you going East in December?

Has any word come from Hagerman?

Were there any noteworthy developments at the hearings on the Tenorio tract?

I have no intimation yet whether I will be wanted in Washington, but the chances are that I will not.

With regards to all the brethren and flock, Dan

Very sincerely yours, George H. H. Baser

Re: Can you read this cursive handwriting? The National Archives wants your help

#156

Post by dahart »

seletskiy wrote: post #109

To tptacek and other guys who seem to have unwavering trust in OCRs/LLMs, as well as to opposite party who think that technology is not there yet — you are all partially right, but somehow fail to hear each other while also spending time on baseless arguing instead of factual examples and attempts to find common truth. Can it be used to greatly simplify efforts by getting through boilerplate? — Yes. Should the result…

Consider using the reply feature so that your comment appears in context.

Also your link goes to the wrong page. Here’s the right one: https://catalog.archives.gov/id/34384201?objectPage=190

Re: Can you read this cursive handwriting? The National Archives wants your help

#157

Post by sandworm101 »

prng2021 wrote: post #137
demosthanos wrote: post #30

Before commenting asking about why they don't just use LLMs, please note that the article specifically calls out that they do, but it's not always a viable solution: > The agency uses artificial intelligence and a technology known as optical character recognition to extract text from historical documents. But these methods don’t always work, and they aren’t always accurate. The document at the top is likely an especi…

Determining whether the latest off the shelf LLMs are good enough should be straight forward because of this: “Some participants have dedicated years of their lives to the program—like Alex Smith, a retiree from Pennsylvania. Over nine years, he transcribed more than 100,000 documents” Have different LLMs transcribe those same documents and compare to see if the human or machine is or accurate and by how much.

This is not an LLM problem. It was solved years ago via OCR. Worldwide, postal services long ago deployed OCR to read handwitten addresses. And there was an entire industry of OCR-based data entry services, much of it translating the chicken scratch of doctor's handwiting on medical forms, long before LLMs were a thing.

Re: Can you read this cursive handwriting? The National Archives wants your help

#158

Post by ivanjermakov »

wkjagt wrote: post #150
ChrisMarshallNY wrote: post #120

They should ask a medical school for help ;) My family is Ivy-League, all the way, and has the worst goddamn cursive writing I've ever seen. It can take me an hour to read a Christmas card from my sister.

I've always wondered how pharmacists can read those prescriptions. There must be some kind of course in university that they followed.

I think with experience they know how each medicine is usually written? It's often easier to listen/read when you already know what it is about.

Re: Can you read this cursive handwriting? The National Archives wants your help

#159

Post by lukeschlather »

sandworm101 wrote: post #157
prng2021 wrote: post #137

Earlier quoted context omitted.

Determining whether the latest off the shelf LLMs are good enough should be straight forward because of this: “Some participants have dedicated years of their lives to the program—like Alex Smith, a retiree from Pennsylvania. Over nine years, he transcribed more than 100,000 documents” Have different LLMs transcribe those same documents and compare to see if the human or machine is or accurate and by how much.

This is not an LLM problem. It was solved years ago via OCR. Worldwide, postal services long ago deployed OCR to read handwitten addresses. And there was an entire industry of OCR-based data entry services, much of it translating the chicken scratch of doctor's handwiting on medical forms, long before LLMs were a thing.

LLMs improve significantly on state of the art OCR. LLMs can do contextual analysis. If I were transcribing these by hand, I would probably feed them through OCR + an LLM, then ask an LLM to compare my transcription to its transcription and comment on any discrepancies. I wouldn't be surprised if I offered minimal improvement over just having the LLM do it though.

Re: Can you read this cursive handwriting? The National Archives wants your help

#160

Post by BugsJustFindMe »

jncfhnb wrote: post #32
BugsJustFindMe wrote: post #8

Earlier quoted context omitted.

> I don’t think I believe that OCR can’t do it but random humans can I do. > OCR is VERY good Uh, my experience is extremely different.

I would challenge you to find a picture of text that you think a human can read and OCR cannot. I’m happy to demonstrate. The text shown in this article is trivial.

Yeah ok, but it might take me a few tries because I don't know what you're using. I hope that's agreeable?

What does your OCR say that these say? The first one isn't too hard for a human (assuming appropriate language skill). The second one is a bit more difficult.

https://imgur.com/a/CDU6Lgs

Post reply on HN