Measurements of generalisation based on information geometry

Neural networks are statistical models and learning rules are estimators. In this paper a theory for measuring generalisation is developed by combining Bayesian decision theory with information geometry. The performance of an estimator is measured by the information divergence between the true distribution and the estimate, averaged over the Bayesian posterior. This unifies the majority of error measures currently in use. The optimal estimators also reveal some intricate interrelationships among information geometry, Banach spaces and sufficient statistics.

Formation and precise geometry control of SNAP microresonators by external electrostatic fields

In SNAP (Surface nanoscale axial photonics) resonators propagation of a slow whispering gallery mode along an optical fiber is controlled by nanoscale variation of the effective radius of the fiber [1]. Similar behavior can be realized in so - called nanobump microresonators in which the introduced variation of the effective radius is asymmetric, i.e. depends on the axial coordinate [2]. The possibilities of realization of such structures “on the fly” in an optical fiber by applying external electrostatic fields to it is discussed in this work. It is shown that local variations in effective radius of the fiber and in its refractive index caused by external electric fields can be large enough to observe SNAP structure - like behavior in an originally flat optical fiber. Theoretical estimations of the introduced refractive index and effective radius changes and results of finite element calculations are presented. Various effects are taken into account: electromechanical (piezoelectricity and electrostriction), electro-optical (Pockels and Kerr effects) and elasto-optical effect. Different initial fibre cross-sections are studied. The aspects of use of linear isotropic (such as silica) and non-linear anisotropic (such as lithium niobate) materials of the fiber are discussed. REFERENCES [1] M. Sumetsky, J. M. Fini, Opt. Exp. 19, 26470 (2011). [2] L. A. Kochkurov, M. Sumetsky, Opt. Lett. 40, 1430 (2015).

The pragmatic motifs of the Jespersen cycle: default, activation, and the history of negation in French

The purpose of this article is to delimit the role of pragmatic specialization in the evolution of negation in French. The change in the marking of sentential negation is believed to proceed in characterized stages that would together constitute the Jespersen cycle. As a marker becomes the default expression of negation, the other markers do not necessarily fade away, and are maintained with specialized roles that include pragmatic functions. One such pragmatic function is that of activation (Dryer 1996), by which a proposition is presented as accessible to the hearer. Activation is shown to motivate the use of preverbal non that competes with 'ne' for several centuries. The claims that the emergence of postverbal pas in early French and the loss of 'ne' in contemporary spoken French are associated with activation are considered on the basis of novel data. It is concluded that pragmatic functions contribute to language change by providing marked options that may be conferred the default status in a grammatical paradigm.

The unnatural history of the lichen Rhizocarpon geographicum

One of the most widely distributed species of crustose lichen is Rhizocarpon geographicum. This unusual organism comprises yellow-green 'areolae' growing on the surface of a non-lichenised hypothallus that extends beyond the margin of the areolae to form a ring. This article describes the general structure of R. geographicum, how the areolae and hypothallus are formed, why the species grows so slowly, and whether it can inhibit its neighbours by releasing allelochemicals.