1
$\begingroup$

I'm currently working on a project that has sensitive data. I'm allowed to see the data on a single platform (company PC), but outside of that platform I cannot process it. I'd like to train a ML algorithm with the data but to do so I'll have to use a different PC. I cannot process the data on that PC though.

Is there anyway to translate and preserve the data in a format that can still be used for ML, but isn't in it's explicit original form?

$\endgroup$
1
$\begingroup$

You can use the following techniques to mask sensitive data:

  • Substitution cipher - any character of plain text from the given fixed set of characters is substituted by some other character from the same set depending on a key.

  • Tokenization masking - mask source string data based on criteria that you specify in an algorithm.

  • Principal Components Analysis (PCA) or other dimension-reducing techniques - combine several features and then carry out ML training only on the resulting PCA vectors.

  • Data Coarsening - decrease the precision or granularity of data to make it more difficult to identify sensitive data within the dataset.

Also, read Data Masking on Wikipedia.

$\endgroup$
0
$\begingroup$

Differential privacy. Where you can inject targeted noise anonymise the data, but you can still make inference.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.