RESEARCH

Here, we want to understand what “seeing” means for a quantum system, and the topic of measurement in quantum mechanics is a world of its own. The question is, with the observations we make of a quantum system (whether it’s a quantum computing scheme or a quantum sensing setup), how can we use this information to perform vision tasks in the classical world? For example, can we use classical shadows to develop generative AI models, or a signal from an extremely faint electromagnetic field to better understand certain biological systems? On the other hand, we are interested in understanding further quantum versions of image processing techniques such as generative models, and to this end we keep an eye on classical-to-quantum data-encoding protocols.

In computer science there are plenty of optimization tasks and projects involving structured databases (for example, at CVC we have a document analysis group where graph-based approaches are prominent). Our goal isn’t so much about implementing algorithms and improving task X, but rather about designing new quantum algorithms for unexplored scenarios.

On the other hand, the idea here is to leverage core QML tasks for building classical-quantum learning models. The overarching question is how quantum correlations can benefit us in performing these tasks—and that’s without even considering the purely quantum version of the concept (a recommender system that outputs a quantum state).

We aim to explore how different RL frameworks can help us both control quantum systems (e.g. configure and re-calibrate quantum devices) and design new ones (e.g. discover new quantum experiments that work under finite resources). On the other hand, we are interested in understanding how certain quantum physics phenomena, such as information storage (or retrieval) capabilities, can assist us in RL tasks. The idea here is to establish connections with the different groups at the CVC working on RL and continual learning (which deals with how models forget what they’ve learned when the environment changes).

We have a strong foothold in transdisciplinary research, involving actors from the humanities, arts, and, of course, science & technology. In fact, we are part of the UAB-Cruïlla Chair, which is a public-private initiative focused on studying AI & Arts in the context of large-scale festivals (beyond performances, organization, etc.), and we are interested in developing further on quantum versions of music production.