Kommentar |
Content:
Basic concepts of information geometry, Riemannian geometry, divergences (statistical and geometric significances), intrinsic geometry of statistical models, exponential and mixture families, generating functionals in statistical physics and Legendre transforms, statistical inference, finite statistical systems, sufficient statistics, application to machine learning, more general geometries.
Format:
- Discussion with one of the advisors before the presentation.
- Presentation manuscript should be ready one week before the talk.
- Second version of manuscript about one week after the talk. |
Literatur |
Main literature:
- Amari and Nagaoka, Methods of Information Geometry (2000)
- Amari, Information Geometry and its Applications (2016)
- Ay, Jost, Lê & Schwachhöfer, Information Geometry (2017)
- Cover & Thomas, Elements of Information Theory (2007) |