Show HN: Person Identification by Keyboard Typing (opens in new tab)

I really don't think that this is a fair appraisal of this. It's a really novel look into how our habits identify us and the creation of it, in itself, doesn't reflect any mal-intent. If you're really so disheartened by the potential use-cases of typing ID like this, you are now aware of it in the first place! You could set out to build a simple tool that randomly adds noise to your keyboard IO and effectively stops this from working. Past that, this is something that others have almost certainly already figured out (be it big tech or government entities) but, even if they hadn't, creation like this is cannot be considered a wrongful act because it's impossible to know the long-term effects of anything before it exists.

1) You should definitely research who Fedor Indutny is.

2) If you think companies aren't already doing this, you're lying to yourself.

3) This is a harmless experiment made to spotlight how simple it is to identify someone just by their typing. This could be used in a multitude of ways.

4) Stop generalizing "people in tech", and stop assuming everyone builds stuff for no reason. This is a deliberate experiment, made on purpose, for a very good reason.

Pretty sure bank apps do this for fraud prevention. Preventing criminals from stealing your money is pretty altruistic

Coursera uses this to make sure exams are given by the same person who registers for the course.

Why would you think "people in tech" are some homogenous group to generalize about?

Unfortunately, tyrants often pay well.

Out of curiosity, did you go to Iowa State? I helped run the study, and there were very good results.

Unfortunately we ran into some mistakes in the communication between DARPA and the university which cut things the project short.

We were also working on mobile identification which included the timing between letter combinations (2-grams if memory serves right) as well as accelerometer data.

This very reasonable sum for a project with funding.

The GradType was creating using my personal funds. Any help is greatly appreciated!

I think it has the potential to be used as a 2FA. However, for this to work - much more training data is required.

British intelligence "authenticated" Russian embassy staff somewhere in the midst of the cold war, and they where using typewriters. So it's way older than 12 years ago. Also: since the authors of this gimmick tool want you to train in by means of typing specifics it implies they made it overcomplicated.

There are multiple companies doing that. BioCatch as a random example. But it does not, for very inherent reasons, work in browser environments. I did not test this classifier, but it'll fail once you switch from chrome to FF.

It’s older than that, Michael Crichton wrote about and posted source code in an Applesoft Basic magazine in early 80s.

The underscore definitely [feels like it] makes me miss most of the commas.

Good point! I've added a small note to explain it. Will re-work the thing later.

This is fixed now.

Very interesting! I don't think that I ever had a sample from a non-standard keyboard layout in the dataset.

While it might sound like this should make your samples very different from others, it could actually act the other way and confuse the network. Hopefully, this will be improved in the next version of it, which I'll start training right after this data collection sprint.

Thank you for taking it to all 5 stars rating. That helps a lot!

Sorry, should be fixed now!

There are also related fields, e.g. identify people based on what they type (not how they type it).

https://stackoverflow.com/questions/4771293/can-an-authors-u...

I'll investigate it. Thank you!

So my keyboard obfuscation usb tool on Kickstarter going to be hit for privacy lovers.

Thx for the link, do you know if it works?

Would they be interested in collaborating with me on this project? My email address is listed on my GitHub profile: https://github.com/indutny/

Thank you!

Most of the collected data is published to github repo: https://github.com/indutny/gradtype/tree/master/datasets

Of course, but with masked identifiers.

Will check it out, thank you!

Similar but I saw two friends show up very early on (and then keep showing up). Maybe people self-select friends based on typing style?

Oh gosh! This is very unexpected. I'm considering removing apostrophes, so this might help.

It is easier than exporting model to TF.js . I wish the tooling was better, but there is way too much custom TF code in gradtype to export it easily.

They'd need to detect and store your keystrokes. If you have JavaScript enabled that's trivial, but hopefully you have JavaScript disabled.

Sorry to hear that! The web page sends each sentence to the server once you type it, but it looks like something has gone wrong.

I'll check server logs to identify the problem. Sorry again!

That's very strange, sorry! What browser/keyboard layout are you using?

This is an ML model: https://github.com/indutny/gradtype/blob/master/src/model.js...

It is trained using (almost) a triplet-loss: http://openaccess.thecvf.com/content_cvpr_2016/papers/Cheng_...

NSA,CIA,TLA have done it since the age of typewriters would be more accurate.

I read about a type of snooping attack already in 2005. https://www.schneier.com/blog/archives/2005/09/snooping_on_t...