2 ms·

I've worked with Biobankers/Bioinformaticians on this type of problem before and none of them wanted us to got the Differential Privacy route. Why? The vast maj

I've worked with Biobankers/Bioinformaticians on this type of problem before and none of them wanted us to got the Differential Privacy route. Why? The vast majority of genomic diseases are rare diseases (with 10s-1000s of carriers). Destroying information is not an option with small sample sizes. For huge GWAS studies, it could be an option. But in the general case, security is preferable over privacy when a choice needs to be made.

alex_hirner 10y ago
Differential Privacy has its limits for small data sets. Luckily, more research is on the way to run arbitrary calculations on multiple datasets that remain private in its entirety (basically homomorphic encryption, but efficiently) [1]. Algorithms that work reasonably well are simple statistical tests and ridge regression [2]. I think these innovations are rare instances where technology can really solves a social problem. [1] https://www.microsoft.com/en-us/research/microsoft-researchers-enable-secure-data-exchange-cloud/ https://www.microsoft.com/en-us/research/microsoft-researche... [2] https://www.microsoft.com/en-us/research/wp-content/uploads/2016/06/SDE.pdf https://www.microsoft.com/en-us/research/wp-content/uploads/...