AIUI the classic procedure for tasks like this is/was:
Hide some graduate students. Tell each to transcribe, and to mark the difficult spots. Give each page to two students. Next, have someone else process the page pairs and resolve conflicts and the marked trouble spots.
As long as one of the students notices that a particular spot is difficult to read, the error is discovered and can be handled by someone who isn't numb from transcribing pages of simple numbers.
The forms are in open access, anyone can fill one form, so they surely did consider variations in accuracy (if not plain sabotage). They are probably submitting the same data to multiple people and the cross-checking the submissions. Also, the totals help to filter out submissions with mistakes. One can also add bound checking for individual values (misplaced decimal point, extra digits etc.). I would love to see how they do it, but they'll probably keep it confidential because trolls and other vandals.
This is exactly how zooniverse works. The first thing you do is go through a few known pieces of data to verify you know how to identify things correctly (more important for galaxy types, etc).
As you start labeling data everything has to reach a consensus. I'm not sure exactly how it works, but it does have multiple people verify each piece of data.
Haha 99% accuracy in weather data would be an amazing improvement. Noisy data and data holes are all we got to work with. Humidity sensors have a memory, storing moist/dry air inside the ceramic.
I've had a look at a few of those and I doubt you'd get 99% accuracy. A lot of the footnotes especially are very hard to decipher. They're also 50 years old, handwriting has changed over time. I'm not sure how well models are trained to read old handwriting, I know that I struggle with it.
For humans, much can be deciphered through context. That's much more complex to do with OCR.
The National Archives in the UK has a research project to try to read > 200 year old handwriting with OCR reliably. They use various tricks like training per-century or per author models which gives a big performance improvement. They also have ways to handle symbols that no longer exist (not in unicode).
Give me a shout if anyone here wants to collaborate.
I work on the visualization and publishing side of this kind of data, but I do know our HTR works even better if we combine authors and centuries into one model, instead of separating. Did you try and compare results?
Maybe they should split in lines each figure, make a coordinates map of the entire figure or so, make a copy and applying a bulk search with a machine for crossing the map and annotate as many undoubtely identifiable numbers as possible. Then paint it in a different color easy to filter or hide it and remember its position.
And then add humans to focuse only in the remaining dificult cases and outliers armed with a reference sample chart. This way an human would need to focus its eyes in 20 characters/image (instead 60 or 100). They would accumulate more completed figures faster and obtain a bigger sense of reward. If the human can't recognize the number, could put a ? in the chain and move on.
Thus the machine would annotate for example 1_3, 4_67_, _8 and the human would write: 5?01 for the same line.
Just an idea, don't know if designing that is easy-peasy or really defiant but the number of possible values is limited in any case, so should be possible to train a machine to recognize some single characters or even entire numbers. Specially if written for the same people.
Is there a technology that would help ocr-ing standardized forms? I’ve got a couple thousand pages of historical train schedules that i would like to digitize (tabulated, printed data, including symbols/icons), but I’m not sure how to automatically recognize the structured data.
I don't know about out-of-the-box ready software. Depending on how much time you have you'd either train your own model with your own ground truth or you can use an open source model. Libraries and archives sometimes make theirs available (I know the Dutch National Archives do). You can train the model and/or HTR the data with software like Transcribus or Kraken (open source).
(disclaimer MS employee but nothing to do with this product)
You could check out the Azure Forms Recognizer. It's designed to do exactly this and make even training a custom model pretty easy. It's also pretty cheap (free up to 500 pages) so you can always do a quick experiment and see if it does what you need it to
I remember getting some reference software and methods from NIST maybe 20 years ago on handwriting recognition. Never put it into practice, but maybe they have some updated software available
I'm curious, what's your interest in the train schedules?
One day at an antique shop, I came across a book from ~1910 which had hundreds of pages of annual reports from railroads with many metrics we'd expect to see in the 10-K reports public companies file.
The book was published annually, but had much of its data in tables with grouped headers and cells, which could make automated OCR-ing with a good (useful) end result challenging.
I think it'd be interesting to map out the Railroad consolidation, track all their financial metrics over time, and do some level of forensic accounting to see if/which companies probably had funny business going on.
I’ve got European rail schedules, which gives an indication of historical travel patterns. Specifically looking into the development of overnight travel.
It would be great if the data entry panel and the main document scrolled separately. Right now I have to scroll up and down to enter data, which is frustrating.
I usually scroll down the main document and use the Tab key to tab between each field to enter data - that way, I don't need to keep scrolling back up and down.
What they should do is make a crypto platform like curecoin but for this where you earn coin for transcription of these documents (and others, could be a cool platform).
Why science must always be done for free, and all the other non scientific stuff, including things dificult to explain logically, are gladly and generously overpaid?
Universities take an economic profit from this volunteer work. They can just share some and hire a few students to do it. And if there is a scarcity of more volunteers, maybe they should think about if repeatedly firing, whimsically dismantling expert teams, and cut the funds to scientists at the slighest opportunity really pays at long term.
Zooniverse was a direct response to a PhD project that Chris Lintott was supervising. His student was tasked with classifying galaxies and in the end decided that the volume of data was just too great.
There is a good supply of volunteers, and the tasks are enjoyed by those who undertake them. Where volunteers identify notable things, they are also credited by name in the resulting papers.
> maybe they should think about if repeatedly firing, whimsically dismantling expert teams, and cut the funds to scientists at the slighest opportunity really pays at long term.
I don't see much evidence for this happening in the UK, where the project is based.
Ha, funny seeing this mentioned on HN, this student was my Professor for Astrophysics during my time in this field. He was a great mentor and taught me (and others) quite a lot, in particular about outreach and science communication. The Zooniverse project is amazing and citizen science in general is a great idea. However, you mention the crux of the issue already: the tasks need to be enjoyable. Looking at galaxies that potentially no other human ever laid eyes on is pretty much as exciting as it may get. A former colleague of mine (in the group of the mentioned student) worked on this interesting structure https://en.wikipedia.org/wiki/Hanny%27s_Voorwerp that was discovered by one citizen scientist in the Zooniverse project (hence the name). It's one of the coolest success stories of crowd-sourced research I've heard about.
> Universities take an economic profit from this volunteer work
This is untrue in many places. Like, not even close. If anything, everyone else economically benefits off the back of groundbreaking and blue-skies research.
For efforts like this, the University really only gets some news recognition blips that people quickly forget.
The PI may have a slightly sexier project if they successfully pull off said endeavor which leads to increased competitiveness in review processes, increased likelihood of funding, and then the University likely gets some overhead in many cases.
The real question for me isn't from Universities making money (which is highly indirect and likely fairly insignificant in volume). The real question is why research is so frequently underfunded and a low priority in this country. There can be more budget for blueskies research but ROI is long, unknown, and research is treated like a business. Research should be treated like an endeavor for knowledge where people doing it need to eat, have a home, etc. and realize it's being done and released to the public for everyone.
Having collected so many data points, I wonder how it was used at the time? Clearly it was probably summed to give yearly totals and possibly monthly averages, and so on. But were the same sheets revisited year after year then?
10+ years ago, Ancestry.com hired people in Asia to transcribe U.S. census returns and vital records. They also used volunteers: http://blogs.ancestry.com/circle/?p=2321
Not sure how it's done now, but I suspect it's very hard to automate owing to inconsistent handwriting styles, unusual names for people and places, footnotes and abbreviations, etc.
Some kind of input validation and user feedback would be welcome. Show a green check mark, if the input annual sum equals the computed sum of the input monthly sums. Typically, all values have two decimal places, thus, validate that there are no more and no less in the user input.
42 comments
[ 3.5 ms ] story [ 96.1 ms ] threadIf I give you 1000 handwritten numbers, do you think you'll make less than 10 mistakes?
Hide some graduate students. Tell each to transcribe, and to mark the difficult spots. Give each page to two students. Next, have someone else process the page pairs and resolve conflicts and the marked trouble spots.
As long as one of the students notices that a particular spot is difficult to read, the error is discovered and can be handled by someone who isn't numb from transcribing pages of simple numbers.
As you start labeling data everything has to reach a consensus. I'm not sure exactly how it works, but it does have multiple people verify each piece of data.
For humans, much can be deciphered through context. That's much more complex to do with OCR.
Maybe they should split in lines each figure, make a coordinates map of the entire figure or so, make a copy and applying a bulk search with a machine for crossing the map and annotate as many undoubtely identifiable numbers as possible. Then paint it in a different color easy to filter or hide it and remember its position.
And then add humans to focuse only in the remaining dificult cases and outliers armed with a reference sample chart. This way an human would need to focus its eyes in 20 characters/image (instead 60 or 100). They would accumulate more completed figures faster and obtain a bigger sense of reward. If the human can't recognize the number, could put a ? in the chain and move on.
Thus the machine would annotate for example 1_3, 4_67_, _8 and the human would write: 5?01 for the same line.
Just an idea, don't know if designing that is easy-peasy or really defiant but the number of possible values is limited in any case, so should be possible to train a machine to recognize some single characters or even entire numbers. Specially if written for the same people.
You could check out the Azure Forms Recognizer. It's designed to do exactly this and make even training a custom model pretty easy. It's also pretty cheap (free up to 500 pages) so you can always do a quick experiment and see if it does what you need it to
https://docs.microsoft.com/en-us/azure/cognitive-services/fo...
One day at an antique shop, I came across a book from ~1910 which had hundreds of pages of annual reports from railroads with many metrics we'd expect to see in the 10-K reports public companies file.
The book was published annually, but had much of its data in tables with grouped headers and cells, which could make automated OCR-ing with a good (useful) end result challenging.
I think it'd be interesting to map out the Railroad consolidation, track all their financial metrics over time, and do some level of forensic accounting to see if/which companies probably had funny business going on.
Universities take an economic profit from this volunteer work. They can just share some and hire a few students to do it. And if there is a scarcity of more volunteers, maybe they should think about if repeatedly firing, whimsically dismantling expert teams, and cut the funds to scientists at the slighest opportunity really pays at long term.
There is a good supply of volunteers, and the tasks are enjoyed by those who undertake them. Where volunteers identify notable things, they are also credited by name in the resulting papers.
> maybe they should think about if repeatedly firing, whimsically dismantling expert teams, and cut the funds to scientists at the slighest opportunity really pays at long term.
I don't see much evidence for this happening in the UK, where the project is based.
Reduction in a 40% of UK funding applicants in the last years:
https://www.bbc.com/news/science-environment-50044659
Lets see how the nice promises will perform at the end
https://www.sciencemag.org/news/2020/03/uk-cues-big-funding-...
This is untrue in many places. Like, not even close. If anything, everyone else economically benefits off the back of groundbreaking and blue-skies research.
The PI may have a slightly sexier project if they successfully pull off said endeavor which leads to increased competitiveness in review processes, increased likelihood of funding, and then the University likely gets some overhead in many cases.
The real question for me isn't from Universities making money (which is highly indirect and likely fairly insignificant in volume). The real question is why research is so frequently underfunded and a low priority in this country. There can be more budget for blueskies research but ROI is long, unknown, and research is treated like a business. Research should be treated like an endeavor for knowledge where people doing it need to eat, have a home, etc. and realize it's being done and released to the public for everyone.
Not sure how it's done now, but I suspect it's very hard to automate owing to inconsistent handwriting styles, unusual names for people and places, footnotes and abbreviations, etc.