Operationally meaningful representations of physical systems in neural networks. Issue 4 (1st December 2022)

• Record Type: Journal Article
• Authors: Poulsen Nautrup, Hendrik; Metger, Tony; Iten, Raban; Jerbi, Sofiene; Trenkwalder, Lea M; Wilming, Henrik; Briegel, Hans J; Renner, Renato
• Is Part Of: Machine learning: science and technology. Volume 3:Issue 4(2022)
• Publication Date: 2022-12-01
• Subjects: representation learning -- neural networks -- reinforcement learning -- Bloch vector -- quantum physics; 006.31
• DOI: 10.1088/2632-2153/ac9ae8
• Languages: English
• ISSNs: 2632-2153

Abstract

Abstract: To make progress in science, we often build abstract representations of physical systems that meaningfully encode information about the systems. Such representations ignore redundant features and treat parameters such as velocity and position separately because they can be useful for making statements about different experimental settings. Here, we capture this notion by formally defining the concept of operationally meaningful representations. We present an autoencoder architecture with attention mechanism that can generate such representations and demonstrate it on examples involving both classical and quantum physics. For instance, our architecture finds a compact representation of an arbitrary two-qubit system that separates local parameters from parameters describing quantum correlations.