Q: 6
Which option describes embeddings in the context of AI?
Options
Discussion
Honestly, D fits best. Embeddings are mainly about creating those dense vector representations so models can work with lower-dimensional data while preserving meaning. C is tempting but that's more about what you *do* with embeddings (like t-SNE/UMAP visualization), not what they fundamentally are. Think D is what AWS wants here, but curious if anyone thinks otherwise.
C/D? I get why people pick C since embeddings help with visualization, but "describe" makes D a better fit. D is more about what an embedding actually is (lower-dim representation), C is more how you might use it. Not 100% sure though, open to pushback.
C , had something like this in a mock and went with C.
Not sure C fits, feels like a trap since visualization isn’t the core of embeddings. D is more about the underlying data representation. So I’d stick with D.
Seen this in a practice test, and D is what they want. Embeddings are all about mapping high-dimensional categorical data into useful lower-dim vectors. The official guide covers this concept too. Fairly confident but open if anyone’s got a different take.
C or D? Embeddings do result in lower-dimensional representations, so D feels right, but sometimes they're visualized, which is C's angle. I'm favoring D since the question says "describes" embeddings, not their downstream use. Agree?
Isn't C possible too if they're focusing on high-dimensional data? Sometimes embedding outputs get visualized, so I might confuse it for that. Not sure if D is the only trap here.
A is wrong, D. The keyword is "representation," not visualization, so D lines up best.
D makes sense, embeddings are for compact numeric data representations (lower dimensions). Official guide and a few practice exams mention this exact phrasing. Pretty sure that's what the exam expects, but let me know if anyone's seen different wording.
If they're asking about describing embeddings, I'd pick C since it mentions high-dimensional data.
Be respectful. No spam.
