Lets say a paper is published which describes a data science algorithm, and the paper is made available on arxiv (no patent or anything else mentioned in paper). The paper is by a university researcher, and the algorithm is not something well known.

A. Is it considered ethical to use my own implementation of that algorithm for:

  1. work/own profit
  2. own learning (I assume yes)
  3. Competition (with monetary gain).

B. On doing so, is it correct that it is the right thing to do to cite the paper (lets say I am required to hand over competition code).

Now lets say the author published some code to go along with the paper under an MIT/other open source license on github, and I mess around with it to fit my needs. What if the paper was by a company (like facebook or google, but does not cite any patent or limitation on use of the code/method) Do the same rules apply or are there other considerations?

I guess I have not found a good blog covering this and any surrounding issues I may not be aware of as self-taught professional human running into this for the first time.

  • 2
    $\begingroup$ Luckily many methods in data science are not encumbered by patents, but some are, for instance RVM are. Also luckily much code is licensed open source, the Apache 2.0 license definitely being your friend when it comes to protection against patent infringement. The question is broader than just patents - in a good way IMO - and it would be nice to see an answer from someone who has to professionally take the ethical and legal considerations of using latest research into account professionally. $\endgroup$ Commented Apr 29, 2018 at 9:50
  • $\begingroup$ My intent was indeed to go broader than just patents. I was trying to use that an example, but also to tease out if there are practices which are 'common but suspect' vs 'more work but more ethical'. The RVM is a good example of something to check for, as there is a scikit package for it, and the github page does not mention the patent, and it appears in the Bishop book which could lead someone to use it w/o due diligence. $\endgroup$ Commented Apr 29, 2018 at 18:55


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.