Quitting PhD 2 years into a 3-year program

Question

I'm currently in a 3-year (European) PhD program in Computer Science/Machine Learning, but I feel like I'm very far behind schedule. I started on a very open-ended project proposal and grant, which I spent my first year reading up on, including taking other courses, TA'ing etc. After the first year, I realized it was a very difficult topic to get anywhere on, and my advisor is also by no means an expert.

For this reason I pivoted to a related subfield of ML that is more mainstream and that I am more comfortable with, and basically started over. Over the past year I've been reading papers, discussing them with people, and trying to get concrete ideas for research. However ideas have been sparse and the ones where I spent time on coding and experiments, they didn't pan out. Now, 2 years into my PhD, I have nothing to show for it. I have no research, no publications, and only some vague ideas on what to work on next. Do I quit? Is it even possible to finish by this point? I'm going to bring this up with my advisor soon, but I would like advice before I do so.

Answer 1

You most probably got into a PhD program with a bad supervisor. This is not your fault. The first option is to talk with your supervisor and see if he thinks there is a chance that you can finish your PhD. Most of the time a PhD is more a social concept than a scientific one: science is hard - if people think that you did things to the best of your ability there is a high chance that they let you graduate with minor results. But be aware that this can set you into a situation of dependence and pressure that can be exploited and might cause mental harm to you. This does depend on your personal relationship to your supervisor. If you think you can go through this period and he is still positive about your projected progress, then stick with it. If not, then start looking for alternatives. There is no shame in starting a second PhD or finding a job in industry. If you openly explain your situation, people will understand. Additionally, doing a PhD is a highly idealistic approach and most people show signs of depression during a situation of hardship. You might want to seek out for additional help if you show signs of depression.

Answer 2

First, be aware that it's extremely common for a "3 years PhD" to actually last more than 3 years, it's very often 4 and sometimes 5 years. The point is that at this stage your case is far from desperate. Since you seem to be motivated to continue, I think the most important thing is to get all the help you can:

• Talk to your advisor. Even if they are "passive", they certainly know how the institution works and what are your chances. You could also ask to have a meeting with the director of studies (or whoever is in charge of the PhD program), it's usually a good idea to let them know in advance that you might need an extension. You might also let them know that it could help to have an additional co-supervisor, if that's possible/desirable. In any case both of them can help you establish a plan for the rest of your PhD, listen to their suggestions carefully.
• In the next 6 months you need to have a clear research path. Talk to any person who could help you, not only your advisor. Go to seminars related to your topic where people discuss ideas. Read papers in the topic you're interested in, not only the best papers from top academics in the field but also average-level publications by PhD students: as a PhD student your aim is not to make a great discovery, just to achieve a piece of decent research work. Try to play your strengths as much as possible: choose a specific topic which suits your interests, your skills and your knowledge.
• Talk to the other PhD students, postdocs and any academic you're comfortable with in your institution. Based on their experience, you should try to know as much as possible about what can happen after the third year: do they know other PhD students who are or have been in a similar situation? What happened to them? How strict is the school about the duration of the PhD, how easy is it to get an extension? Very often the official position is that PhDs must be finished within 3 years, but in practice there are many exceptions; you can only know about this by asking people.
• I assume you have a PhD funding for 3 years right? Your PhD is likely to last longer, so you need to start planning for that: is there any kind of grant, teaching position that you could have after the 3rd year? In the worst case scenario, do you have savings and how long would you last on them?

As I said I'm assuming that you want to continue, but there's no shame leaving if you don't feel interested anymore. It's not too hard to find a good job with a CS/ML background nowadays. The one thing I really don't recommend is to quit in the hope to restart a new PhD program.