3

I have written a thesis about applying some AI techniques to a simulation that my professor introduced. To make this happen, I had to implement a very large amount of software, because the simulation works in another language, follows different principles etc.

After 6 months of hard work, I did not reach my goal. I was not able to actually implement the AI technique I wished to apply to the problem. I created a lot of reusable components that future developers (and in fact other students as I was told) will want to use. But the actual question and also the title of the thesis (which both had to be fixed ahead of time and cannot be changed once it has been started) don't match the main body of my work.

Can a thesis still be considered good even if the question was not answered because it was not "reached"? If not, why? If yes, why? I can imagine this going either way but I'd like to hear opinions and experiences.

Basically I would require another 1-2 months and maybe someone to work on the problem as a team to get new ideas. I believe I still worked in a scientific manor (I prioritized reusability and usefullness over just reaching the question but without it being easily reusable) and no one ever said science is only good if it never misses a deadline. But I can also imagine an argument that simply states an unanswered research question is a bad work.

My original intention was to apply this AI technique to the common example research problems (like Atari games or locomotion) but my professor insisted I apply it to his field of work (energy market simulation). I had to agree if I wanted him to take my thesis and I didn't know how complex the mapping of his problem to the commonly used tools would be. The argument "a good researcher is also capable of creating a reasonable problem scope" therefore is a little unfair as he sort of forced me to extend it.

Alternative B I write a preface / authors comment at the beginning, stating that the title/research question won't be matched but due to university regulations may not be changed retrospectively.

Alternative C: A fellow researcher recommended me to just change the question and content but ignore the title. I hand in my thesis under the forced title but publish it under a more suitable title. That seems a little "fake" as I don't want to lie about my original goal and me missing the final goal.


Edit: Most suggestions go along the lines of "depends on your professor or institution". While I fear this is probably the only right answer that helps me personally, it doesn't seem right that it depends on the professors attitude or personal opinion. From an abstract perspective, is it OK to rate a scientific work as a bad work if it fails to reach a previously defined target when new information came to light along the way? Probably not. But is it common practice to change the research question at the end to better match / suit the line of argumentation? It seems to me the question should reflect what the researcher asked him/herself when he/she started the inquiry. If the results are not what was expected, that doesn't mean it's bad results.

  • Seem to have read this question the other day - and answered it.... – Solar Mike Jul 12 '18 at 19:04
  • Your school forces you to write about exactly the question that was given out when you first started?!? Idiocracy knows no limits. – Karl Jul 12 '18 at 19:43
  • @Karl I may be able to change the question (I am still waiting for an answer of my professor but it has been waiting 2 weeks now despite several emails) but the title is fixed and with 3 weeks left, it's tough to restructure the whole line of arguments now anyways. – pascalwhoop Jul 12 '18 at 20:12
  • Have you asked your professor what you should do? The answer to this question is: expectations vary. – Dawn Jul 12 '18 at 20:16
  • @Dawn he doesn't respond to my inquiries anymore. I could contact my other contact person but she usually just refers to him as the authority that must make the decision. – pascalwhoop Jul 12 '18 at 20:23
4

The short answer to your question is yes, you can have a good result even if you don't (completely) answer the original question. However, the advice you really need is from your professor. Don't express negative feelings about how he "forced" you or anything like that. Ask him for his analysis of your work and how to best position it.

Many research questions take a long time to resolve. Don't be discouraged by that.

But the work isn't a failure, and you, especially, shouldn't think of it that way. You want to ask yourself, however, "What, exactly, did I demonstrate in this work". Your wording of the rest should revolve around answering that question and showing how you did a good job of it.

You can have a final section that discusses how this work is part of a larger work and briefly describe the problem you started out to solve.

However, some of your statements confuse me. You seem to be saying that at some earlier point you gave a title to the university that makes a claim of success to the earlier question, and that that can't be revised. Such a policy would surprise me and I would make doubly certain that you are correct. But advice from your professor would also be valuable here.

When it comes time to publish your work, there is no rule anywhere that I know of that restricts you in the title of the paper(s) that get produced. So even if you are sort of "stuck" with an inappropriate title, you can just treat it long term as an anomaly. You then have a little joke you can share with future students if you stay in academia.

1

Yes, it can. Research is too flexible to be sure about a year in advance. Perhaps you found out that the question is the wrong one to ask, or perhaps someone proved that a different method is superior - it'd be foolish to persist on the same path then.

In any case, talk to your professor, who will know what your institution's policies are.

  • while my professor didn't answer, his assistant who is supposed to be my contact person did. She told me to just stick with my current trajectory and discuss the shortcomings in the conclusion and explain what stopped my from completing the question. I guess the short answer would be "time" and the long answer will be some pretext why I couldn't develop software faster in the 6 months to get to the goal in time. – pascalwhoop Jul 14 '18 at 8:27
0

I found the answer I gave to the other question :

You can have a valid result showing that the "XXX method" is not the best in YYY situation - that is, or can be, useful work for someone who does not waste their time on that possibility.

They can then focus on a different direction - which you can also do ie checking out if the Tabu search works for ZZ or AA for example..

  • The issue is not that I wanted to try to apply XXX on YYY but it didn't work. I wanted to apply XXX to YYY but the professor insisted it would be ZZZ for which it turned out I had to implement A, B, C, and D. I implemented A - D but then time was gone and I was not able to apply XXX to ZZZ anymore. – pascalwhoop Jul 12 '18 at 20:13
  • Then that was your problem - you should not apply XXX "ON" YYY but you shoud apply XXX "TO" YYY.... – Solar Mike Jul 12 '18 at 20:25

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.