Providing data of unpublished thesis

Question

As part of my thesis in an ML/NLP I build some datasets (corpus and word embeddings) and of course the thesis.

After the first revision round, the lectors suggested to cut many of the topics discussed and focus on just three.

My concern is that now my director is requesting me to give my data/models to other student, so that she can make her thesis in a related topic. Even more, he suggested her to do exactly what I told him that I'm doing for an article based on the discarted parts of my thesis.

At this point I have no thesis defense date, almost no feedback from my director on the document (except changing the colors on the figures) and a deadline to complete the program at midyear, otherwise I get expelled. Although, the document was ready since January and I was expected to do the defense in April.

Should I provide the pretrained models? I should raise my concerns to someone else at my university?

Answer 1

I would attempt to raise concerns first with your own thesis advisor/director. Sometimes it is worthwhile to just speak directly to those with whom we have disagreements.

Does your department have a graduate coordinator or a professor overseeing the graduate on the whole? He or she may be the next point of contact.

As for the discarded portion of your thesis, it may be okay to allow another student to join in on that work. (Although, just giving it over to them wholesale is slightly different). Even if it goes into the other student's thesis, nothing (in theory) prohibits the both of you from publishing a paper jointly on the work.

Note that many advisors see any work you did under their supervision (as minimal as that supervision may have been) as belonging to them themselves. It might have been work that you performed, but if your advisor funded it or gave any input (or even existed on the planet) while you did that work, he may see that work as being his to make decisions on.