Quantum Unsupervised?

Good morning Pennylane’s team.
Quick question: is there anything about Quantum Unsupervised ML algorithms? For example the quantum version of classical clustering algorithm? If someone knows about previous works with the implementation could post here?
Thanks in advance.

Hi @checcopo,

Indeed, you have the unsupervised learning approach, too, in quantum. The best example is Quantum K-means.

Hey @checcopo,

You can also check out this demo:

It’s got unsupervised QML :slight_smile:. But there’s also a ton of research on this. A quick Google search shows a lot of hits!

Thank you @Christophe_Pere and @isaacdevlugt for answering. Probably I am a little bit confused about this topic. I saw on the web that it’s possible to make protocols to cluster data by using Qiskit as many examples are based on it. My question is, is it possible to use Pennylane and, in that case, train a PQC to cluster data? I mean, in Kmeans algorithm you assign k number of centroids randomly in your data-space and, according to a condition on the distance between each centroids and the data, you build clusters. Is it possible to update the parameters of a PQC to realize this process or it is simple no sense?
Thanks in advance, sorry for being so confused.

Hey @checcopo, this will most certainly be possible in PennyLane! Do you have examples in Qiskit that you’re familiar with that you can share? We have some great PennyLane-Qiskit features (see our plugin: PennyLane-Qiskit Plugin — PennyLane-Qiskit 0.35.1 documentation), but I’m 99.999% certain you can translate any Qiskit example to pure PennyLane :slight_smile: