Abstract: Thermodynamic laws have been found applicable to many systems within the universe. But are they applicable to the universe itself as a whole system? Based on the assumption that they are, we are able to propose a modality rooted in quantum physics which can permit astronomical increases in the entropy, phase-space volume and thus the number of position and momentum coordinates that are available in a system, in spite of any prevailing adiabatic conditions. We conclude that a study of the thermodynamic consequences of energy introduction into a state at low or absolute zero temperature may increase our understanding of any possible cosmic genesis.

Abstract: We define an entropy based on a chosen governing probability distribution. If a certain kind of measurements follow such a distribution it also gives us a suitable scale to study it with. This scale will appear as a link function that is applied to the measurements. A link function can also be used to define an alternative structure on a set. We will see that generalized entropies are equivalent to using a different scale for the phenomenon that is studied compared to the scale the measurements arrive on. An extensive measurement scale is here a scale for which measurements fulfill a memoryless property. We conclude that the alternative algebraic structure defined by the link function must be used if we continue to work on the original scale. We derive Tsallis entropy by using a generalized log-logistic governing distribution. Typical applications of Tsallis entropy are related to phenomena with power-law behaviour.

Type of the Paper: Full Research Paper
Title: Additive Composed Quantum Statistical Entropy
Authors: Philipp Dedié, Wolfgang Muschik
Abstract: The incompatibility between the quantum statistical description of isolated undecomposed systems by the subadditive VON NEUMANN entropy and the irreversible behavior of the corresponding subdivided two-part composed system is discussed. An entropy definition for the composed system is offered. This composed quantum statistical entropy is additive and describes in contrast to the VON NEUMANN entropy the irreversibility of the composed system. This entropy definition dissolves another incompatibility, namely the one using the VON NEUMANN entropy for any isolated undecomposed system and the thermodynamic axiom testifying that any reversible composed system can only consist of reversible subsystems. Moreover, the proposed composed entropy definition yields an endoreversible description of the composed system in consideration.
 
Manuscript ID: Entropy-12-10
Type of the Paper: Full Research Paper
Title: Maximum Entropy Parameter Learning at Elevated Training Temperature
Authors: Ronny Melz
Abstract: ForMaximum Entropy (ME) parameter inference, the Improved Iterative Scaling algorithm (IIS) is often preferred over Generalized Iterative Scaling (GIS) due to its better convergence properties. But effectively, IIS requires the feature sum for each training event to be drawn from quite a limited, finite set of discrete values to allow for an efficiently computable parameter update step. Quite some generality of ME models is lost by requiring the feature functions to sum to discrete values. We re-interpret the maximum feature sum (originally determined by the GIS convergence proof) as an inverse “training temperature”, i.e. an additional free hyper parameter of the model. We provide empirical evidence that GIS outperforms IIS for suitable values of the training temperature, especially in the most interesting early iterations, despite its less complex implementation.
 
Manuscript ID: Entropy-12-11
Type of the Paper: Full Research Paper
Title: Graph Entropy and Conditioning
Authors: Arthur Ramer and Marian Grendar
Abstract: How to perform conditioning when certain letters/outcomes are not distinguishable? Distinguishability being specified by a graph, we apply K¨orner’s graph entropy and related information divergence on graphs to address this question.