SOP strategy: ML or not ML?

[Schweinchen]

I am applying to PhD programs in Stat and CS (and a few in applied math and EE) looking to do machine learning. Somebody I know who's on the admissions committee at a top school told me that when it comes to my SOP, it's better not to emphasize too much that I want to do ML since they get a ton of applicants who write that, and that on the committee they tend to lump applicants into different groups so that if you get assigned to the ML group, you face much stiffer competition. She said that in reality a lot of applicants say that they want to do something else and after they got admitted they switched to ML.

The problem is that I've already spent a lot of time writing my SOP, doing just what she told me not to. In fact since I've been gearing towards a PhD in ML, all my experience (research projects, courses, internships, etc) are somewhat ML-related. It never occurred to me that this could be a bad strategy. If I remove the focus on ML, my SOP will be weaker. Besides it's kind of hard to hide since my two of my letters will be from ML professors and my entire profile clearly points towards ML.

Anyone else who's facing a similar dilemma?

[insert_name_here]

You're overthinking it, just stick with ML. The difference between you having a research area you're excited+prepared for (ML), as opposed to not having a well-focused background is going to outweigh any of the effects you describe (if they even exist). FWIW, I've observed the opposite - lots of people put ML in their SOP and end up doing other things.

[Schweinchen]

That's what I thought too but great to have independent confirmation. Thank you, pal.