We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Entangled q-Convolutional Neural Nets

Abstract: We introduce a machine learning model, the q-CNN model, sharing key features with convolutional neural networks and admitting a tensor network description. As examples, we apply q-CNN to the MNIST and Fashion MNIST classification tasks. We explain how the network associates a quantum state to each classification label, and study the entanglement structure of these network states. In both our experiments on the MNIST and Fashion-MNIST datasets, we observe a distinct increase in both the left/right as well as the up/down bipartition entanglement entropy during training as the network learns the fine features of the data. More generally, we observe a universal negative correlation between the value of the entanglement entropy and the value of the cost function, suggesting that the network needs to learn the entanglement structure in order the perform the task accurately. This supports the possibility of exploiting the entanglement structure as a guide to design the machine learning algorithm suitable for given tasks.
Subjects: Machine Learning (cs.LG); Quantum Physics (quant-ph); Machine Learning (stat.ML)
Cite as: arXiv:2103.11785 [cs.LG]
  (or arXiv:2103.11785v1 [cs.LG] for this version)

Submission history

From: Miranda C. N. Cheng [view email]
[v1] Sat, 6 Mar 2021 02:35:52 GMT (247kb,D)

Link back to: arXiv, form interface, contact.