Are HMMs still used for unsupervised tasks? [D]
Heat trend
The percentage is based on available heat signal, not comment count or independent people.
The user is exploring Hidden Markov Models (HMMs) as a baseline for unsupervised "dataset exploration/discovery" with unstructured, unannotated data. They are seeking to understand if more modern approaches, including deep learning-based methods, have entirely replaced HMMs for these types of tasks, suggesting a need for insights into data structure and semantics.
I'm exploring Hidden Markov Models (HMMs) as a baseline method for "dataset exploration/discovery" where I have a bunch of unstructured data with no annotations, and wish to gain insights about the structure and semantics of the data within. I was wondering if there are more modern (deep learning based or otherwise) approaches which have completely superseded HMMs for such tasks.