Program

October, 16

13:45 – 14:00

Reception

14:00 – 14:30

Amos Golan (Info-Metrics Institute, American University; Santa Fe Institute; Pembroke College, Oxford)

Foundations of Info-Metrics in Pictures

Info-metrics is the science and practice of modeling, reasoning, and drawing inferences under conditions of noisy and insufficient information. It is at the intersection of information theory, statistical inference, and decision-making under uncertainty. It plays an important role in helping make informed decisions even when there is inadequate or incomplete information because it provides a framework to process available information with minimal reliance on assumptions that cannot be validated.

In my talk, I will very briefly introduce the basic idea, motivation and framework using graphical representations of the problem and the theory behind the solution. I will also provide graphical representations of a number of case studies from across the scientific spectrum.

14:30 – 15:00

José P. Sande Lemos (CENTRA, Physics Department, Instituto Superior Ténico, University of Lisbon)

Black Hole Entropy and Fundamental Degrees of Freedom

Black holes are at the center stage of fundamental physics. From their generic properties, one can infer important clues to what a fundamental theory, a theory that includes gravitation and quantum mechanics, should give. Here we review the classical properties of black holes and their associated event horizons, as well as the quantum and thermodynamic properties, such as the temperature, derived from the Hawking radiation, and the entropy. Then, using the black hole properties we discuss a universal bound on the entropy for any object, or for any given region of spacetime. We then show that associating entropy with the number of quantum degrees of freedom, via statistical physics, leads to a holographic principle, with the conclusion that the degrees of freedom of a given region are in the area A bounding the region, rather than its volume V, as naively could be thought. Surely, a fundamental theory has to take this into consideration.

15:00 – 15:30

Paulo Ferreira (IEETA, University of Aveiro)

Entropy, Complexity and Applications to the Measurement of Similarity

The concepts of entropy and algorithmic complexity are interesting on their own but can also be applied to solve large-scale classification problems in somewhat surprising ways.

The goal of this talk is to discuss how these fundamental concepts can be combined with probabilistic models and data compression techniques to detect in multiple streams of data regions of similarity, regions of uniqueness and rearrangements. The applications include authorship attribution and genomic data analysis.

15:30 – 16:00

Coffee break

16:00 – 16:30

Elvira Silva (CEF.UP, University of Porto)

Maximum Entropy and Regulation in the Energy Sector

The evidence from the literature on incentive-based regulation in the electricity sector indicates that the size of a country’s electricity sector (i.e., the number of companies in the electricity value chain) tends to influence the choice of benchmarking methods, as well as the specification of the frontier model. The size of a country’s electricity sector and the problems with (or lack of) data are among the reasons pointed out by national regulators who do not employ benchmarking techniques. Countries with a small number of transmission and distribution companies are constrained in the choice of methods, because the limited number of similar companies is a barrier to the use of frontier benchmarking methods. A regulator rarely regulates a large number of similar firms; firms usually vary in size and other characteristics (e.g., shareholding composition or structure, age of assets, portfolio). Some national regulators have been facing a problem of ill-posed frontier models. The main purpose of my talk is to present an alternative benchmarking approach (to the commonly used DEA) based on Stochastic Frontier Analysis with the generalized maximum entropy estimator, to set price controls within incentive-based regulation.

16:30 – 17:00

Andreia Dionísio (CEFAGE, University of Évora)

GME Versus OLS - Is It Worth the Effort?

In this communication we intend to show the main advantages (and limitations) of GME facing to OLS procedures to estimate regression models in specific conditions. In order to better show our purposes, we will present two examples: (i) utility function estimation; and (ii) evaluation of financial integration in the European Union.

We compare the performance of the GME estimator with ordinary least square (OLS) in a real data sample setup. The results confirm the ones obtained for samples through Monte Carlo simulations. The difference between the two estimators is small and it decreases as the width of the parameter support vector increases. Moreover, the GME estimator is more precise than the OLS one. Overall, the results suggest that GME is an interesting alternative to OLS in the estimation of regression models, despite the difficulties in computational terms. So, it is worth the effort?

17:00 – 17:30

Maria Conceição Costa (CIDMA, University of Aveiro)

Maximum Entropy in Inhomogeneous Large-scale Data

Large-scale data refers to datasets that are large in different ways: there are many observations, many variables, or both, or data is recorded in different time regimes or taken from multiple sources. Another difficult issue that is usually related to large-scale data is the presence of inhomogeneity: data is neither independent, identically distributed, nor stationary. In this context, obtaining reasonably good statistical properties with a computationally efficient analysis becomes a challenge. In this work, the concepts of Info-Metrics are introduced to the analysis of inhomogeneous large-scale data and the framework of information-theoretic estimation methods is briefly presented. Sub-sampling and a new aggregation procedure based on the normalized entropy are proposed. A simulation study is presented and the preliminary results obtained clearly indicate that normalized entropy methods provide very satisfactory solutions.

October, 17

09:00 – 10:30

Amos Golan (Info-Metrics Institute, American University; Santa Fe Institute; Pembroke College, Oxford)
Foundations of Info-Metrics — A Tutorial (Information-Theoretic Methods of Inference) (part I)

10:30 – 11:00

Coffee break

11:00 – 12:30

Amos Golan (Info-Metrics Institute, American University; Santa Fe Institute; Pembroke College, Oxford)
Foundations of Info-Metrics — A Tutorial (Information-Theoretic Methods of Inference) (part II)

Note: This tutorial will be based on the book “Foundations of Info-Metrics: Modelling, Inference, and Imperfect Information“.

The support page: http://www.info-metrics.org

 Download Program / PDF (228KB)

Organizing Committee

Andreia Dionísio, CEFAGE, University of Évora

 

Maria Conceição Costa, CIDMA, University of Aveiro

Elvira Silva, CEF.UP, University of Porto

 

Pedro Macedo, CIDMA, University of Aveiro

Sponsors

This workshop is supported in part by the Portuguese Foundation for Science and Technology (FCT – Fundação para a Ciência e Tecnologia), through CIDMA – Center for Research and Development in Mathematics and Applications, within project UID/MAT/04106/2013.