Quantum Hamiltonian embeddings proposed for supervised learning
Original title: Supervised Learning with Quantum Hamiltonian Embeddings
A dataset is encoded into one Hamiltonian whose ground state classifies; the method was demonstrated on IBM Heron processors.
Researchers introduce quantum Hamiltonian embeddings, a supervised framework that encodes a classical training set into one parameterized Hamiltonian. Training tunes that Hamiltonian, and its approximate ground state defines the classifier, with inference possible from a classical or quantum state representation. They report efficient gradient estimation, generalization guarantees, and, under standard cryptographic assumptions, hypotheses beyond efficient classical algorithms, plus a hardware demonstration on IBM Heron processors using a standard benchmark dataset.
- Inference can use a classical or quantum representation of the learned state.
- The paper proves favorable generalization guarantees and efficient gradient estimation.
- Cryptographic assumptions separate some hypotheses from efficient classical algorithms.
- The framework was demonstrated at scale on IBM Heron processors.
From the original article
1 min read at arxiv.orgMachine learning is widely regarded as a promising application in which quantum computing could offer a computational advantage. However, general quantum machine learning models with provable guarantees on performance, trainability, and efficient inference remain limited.
By Riccardo Molteni · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.