2

I have been working on an NLP project (Natural Language Processing) regarding a variety of Arabic. I had to observe and look for patterns and rules in this variety as the resources are almost absent or very rare. I then started working on this module that classifies, tokenizes, stems, and returns the part-of-speech of each instance of the input. I'm afraid of the high possibility that one of my supervisors with such a history of stealing others work and labeling it as his. When I present my research, must I present the source code to the supervisor, or must I only live-experiment the model? I do not have a problem with distributing or open-sourcing the whole thing; I am willing to. But I am afraid he takes my work and make it his with no credits or rights.

Noah D
  • 68
  • 4
coredumped0x
  • 123
  • 3
  • 6
    Why do you mistrust your supervisor? – Buffy May 07 '19 at 15:28
  • 1
    Closely related: https://academia.stackexchange.com/questions/8278/worry-about-stealing-of-research-ideas?r=SearchResults – henning May 07 '19 at 15:33
  • 9
    Why are you working with a supervisor you trust so little? – Bryan Krause May 07 '19 at 16:05
  • If you do not mind to open source it or GPL/MIT/Apache license it, why not make it available on Github or similar? – Captain Emacs May 07 '19 at 17:56
  • I think it's not really a wise action to make it available before you published it somewhere if this algorithm plays a key role in your research. At least try to publish a description + application of this algorithm in a preprint server and then make the code public and explicitly say anyone that wants to use your code should give proper reference to your repository as well as preprint. – Mithridates the Great May 07 '19 at 18:03
  • @AloneProgrammer True, but his problem is that he will have a problem publishing it without his supervisors, especially if one of them is someone who "nostrifies" credit. – Captain Emacs May 07 '19 at 18:07
  • @CaptainEmacs As far as I know the owner of his work (e.g. code or algorithm) is the university not the advisor. But, it's a bit vague I guess, at least in my university guidelines. For example, if the owner is the university, am I allowed to prevent my advisor's access to the actual code? I don't think so, cause if someone is supposed to be your co-author (e.g. OP's advisor) should be able to understand and reproduce the results and it makes sense for his advisor to request an access to the code. I think it pretty much depends on advisor's ethics to not steal work of young researchers... – Mithridates the Great May 07 '19 at 18:14
  • Yes, they can ask access (or just use it), but author's name is already there. Of course, author should take care that they do cite the supervisor in the code if they use their ideas. Some students think they invented something when it turns out that they profited from the supervisors background, direction, advice, and the context of the supervisors' group. I am afraid, while there is a downhill power gradient superviser-supervisee, the integrity gradient's direction is not always as clear. – Captain Emacs May 07 '19 at 19:10
  • @Buffy. I do not mistrust him. I'm just labeling some of his history with copyright infringement. It's not the first time he's even done it. Thank you. – coredumped0x May 08 '19 at 17:38
  • @Bryan, I would love a supervisor who I could trust, but you could say I had no saying in that. Excluding the field of reseach criteria, we were assigned to supervisors "automatically & randomly". Thank you all for your time and help. Very appreciated. – coredumped0x May 08 '19 at 17:38

1 Answers1

5

The best way to protect your research from someone else taking credit is to publish it first. Even then, someone with a stronger reputation or more recognition can attempt to take credit for it. This is a risk you take in research and which is made worse by working with a supervisor "with such a history of stealing others work."

The more public your contributions are, the less likely someone else can steal it from you. If you have multiple supervisors make sure they are all aware of your contributions. Maybe talk with them about finding a chance for you to present at a poster session, conference, or research forum. Their names will likely be on the project (and probably deserve to be) if they are your supervisors, but yours will too.

Noah D
  • 68
  • 4
  • 2
    In particular, publishing solves the copyright problem. It is yours to hold. However, if you have something that is patentable, you need to take forward action to obtain the patent. Unlike copyright, patent doesn't happen automatically. – Buffy May 07 '19 at 18:00