It's important to mention that this is independent of being remote.
If you want to have user stories and mock-ups, start spending your time making those yourself and push them back to whoever is giving you the requirements for comment.
If they question the need for you to be spending time doing that, it's an opportunity to ask for more specific direction on what you should be spending your time doing. Chances are it's a classic case of not knowing what they want, but knowing what they don't want, so by having their requirements reflected back to them they'll come back with "Ah, but what I meant was..." and you'll get more clear requirements.
If it's still murky after that and you really can't get more concrete requirements than "We need this magic box made" then consider that some places genuinely just aren't good places to work, and consider this a learning opportunity and point for future comparison.
If (1), you're probably best served not asking for much help. If (2), you need to educate & manage expectations about the complexity and necessary steps for the feature. Either case, you would probably be best served by prioritizing the input/output functionality (input a picture, output a prediction) over optimizing precision/recall on the algo. This allows you to 1) show progress to others and 2) you can start having conversations about false positives/negatives by showing different images and the prediction output.
Having worked remotely for a number of years, and having run a startup, I think this is the root cause of your issue. You're not going to get well defined requirements because the definition of "well defined" varies from person to person, and because startups generally don't plan features in great detail (tbh, as a rule established businesses don't either).
If you want to make a bigger impact, start being more proactive in asking what features need to do in order to be accepted. If you don't know something, think about it, then write a clear document that sets out what you think you need to do and points to specific things that you need clarification about. That's what I'd expect any developer to do, regardless of whether they're an intern or an experienced lead developer. The key thing no one in a business is simply there to execute instructions. You have to think for yourself.
tl;dr If you don't know what you're supposed to be building there's no way you can build a thing that works, so think through the problem and ask good questions.