BEGIN:VCALENDAR VERSION:2.0 PRODID:researchseminars.org CALSCALE:GREGORIAN X-WR-CALNAME:researchseminars.org BEGIN:VEVENT SUMMARY:Maciej Koch-Janusz (University of Zurich) DTSTART:20210222T170000Z DTEND:20210222T180000Z DTSTAMP:20260423T021426Z UID:QM3/33 DESCRIPTION:Title: Sta tistical physics through the lens of real-space mutual information\nby Maciej Koch-Janusz (University of Zurich) as part of Quantum Matter meets Maths (IST\, Lisbon)\n\n\nAbstract\nIdentifying the relevant coarse-grain ed degrees of freedom in a complex physical system is a key stage in devel oping effective theories. The renormalization group (RG) provides a framew ork for this task\, but its practical execution in unfamiliar systems is f raught with ad hoc choices. Machine learning approaches\, on the other han d\, though promising\, often lack formal interpretability: it is unclear w hat relation\, if any\, the architecture- and training-dependent learned " relevant" features bear to standard objects of physical theory.\n\nI will present recent results addressing both issues. We develop a fast algorithm \, the RSMI-NE\, employing state-of-art results in machine-learning-based estimation of information-theoretic quantities to construct the optimal co arse-graining. We use it to develop a new approach to identifying the most relevant field theory operators describing a statistical system\, which w e validate on the example of interacting dimer model. I will also discuss formal results underlying the method: we establish equivalence between the information-theoretic notion of relevance defined in the Information Bott leneck (IB) formalism of compression theory\, and the field-theoretic rele vance of the RG. We show analytically that for statistical physical system s the "relevant" degrees of freedom found using IB compression indeed corr espond to operators with the lowest scaling dimensions\, providing a dicti onary connecting two distinct theoretical toolboxes.\n LOCATION:https://researchseminars.org/talk/QM3/33/ END:VEVENT END:VCALENDAR