Strong minimax lower bounds for learning

Minimax lower bounds for concept learning state, for example, thatfor each sample size $n$ and learning rule $g_n$, there exists a distributionof the observation $X$ and a concept $C$ to be learnt such that the expectederror of $g_n$ is at least a constant times $V/n$, where $V$ is the VC dimensionof the concept class. However, these bounds do not tell anything about therate of decrease of the error for a {\sl fixed} distribution--concept pair.\\In this paper we investigate minimax lower bounds in such a--stronger--sense.We show that for several natural $k$--parameter concept classes, includingthe class of linear halfspaces, the class of balls, the class of polyhedrawith a certain number of faces, and a class of neural networks, for any{\sl sequence} of learning rules $\{g_n\}$, there exists a fixed distributionof $X$ and a fixed concept $C$ such that the expected error is larger thana constant times $k/n$ for {\sl infinitely many n}. We also obtain suchstrong minimax lower bounds for the tail distribution of the probabilityof error, which extend the corresponding minimax lower bounds.

From correspondence analysis to multiple and joint correspondence analysis

The generalization of simple (two-variable) correspondence analysis to more than two categorical variables, commonly referred to as multiple correspondence analysis, is neither obvious nor well-defined. We present two alternative ways of generalizing correspondence analysis, one based on the quantification of the variables and intercorrelation relationships, and the other based on the geometric ideas of simple correspondence analysis. We propose a version of multiple correspondence analysis, with adjusted principal inertias, as the method of choice for the geometric definition, since it contains simple correspondence analysis as an exact special case, which is not the situation of the standard generalizations. We also clarify the issue of supplementary point representation and the properties of joint correspondence analysis, a method that visualizes all two-way relationships between the variables. The methodology is illustrated using data on attitudes to science from the International Social Survey Program on Environment in 1993.

The choice of the form of representation in multinational banking: Evidence from Spain

This paper investigates foreign direct investment in the bankingsector. The attention has been addressed to test the importanceof OLI advantages as the determinants of the bank's decisionto invest in foreign locations. Nevertheless, since banks canexpand their activities abroad through different organizationalforms that imply different levels of foreign involvement, theissue of the form of representation has been tackled. Theresults show the importance of OLI advantages in the form ofrepresentation in multinational banking.

Forward induction and the excess capacity puzzle: An experimental investigation

While the theoretical industrial organization literature has long arguedthat excess capacity can be used to deter entry into markets, there islittle empirical evidence that incumbent firms effectively behave in thisway. Bagwell and Ramey (1996) propose a game with a specific sequence ofmoves and partially-recoverable capacity costs in which forward inductionprovides a theoretical rationalization for firm behavior in the field. Weconduct an experiment with a game inspired by their work. In our data theincumbent tends to keep the market, in contrast to what the forwardinduction argument of Bagwell and Ramey would suggest. The results indicatethat players perceive that the first mover has an advantage without havingto pre-commit capacity. In our game, evolution and learning do not driveout this perception. We back these claims with data analysis, atheoretical framework for dynamics, and simulation results.

Insurance and safety after September 11, 2001: Coming to grips with the costs and threats of terrorism

This chapter, originally written as a consequence of the terrorist attacksof September 11, 2001, provides an elementary, everyday introduction tothe concepts of risk and insurance. Conceptually, risk has two dimensions:a potential loss, and the chance of that loss being realized. People can,however, transfer risk to insurance companies against the payment ofso-called premiums. In practice, however, one needs accurate assessmentsof both losses and probabilities to judge whether premiums are appropriate.For many risks, this poses little problem (e.g., life insurance); however,it is difficult to assess risks of many other kinds of events such as actsof terrorism. It is emphasized, that through evolution and learning, peopleare able to handle many of the common risks that they face in life. Butwhen people lack experience (e.g., new technologies, threats of terrorism),risk can only be assessed through imagination. Not surprisingly, insurancecompanies demand high prices when risks are poorly understood. In particular,the cost of insurance against possible acts of terrorism soared afterSeptember 11. How should people approach risk after the events of that day?Clearly, the world needs to protect itself from the acts of terrorists andother disturbed individuals. However, it is also important to address the root causes of such antisocial movements. It is, therefore, suggested thatprograms addressed at combatting ignorance, prejudice, and socialinequalities may be more effective premiums for reducing the risk ofterrosrtism than has been recognized to date.

Fact Sheet: House and Building Fires, 2009

A fire can engulf a structure in a matter of minutes. Understanding the basic characteristics of fire and learning the proper safety practices can be the key to surviving a house or building fire.

Fact Sheet: House and Building Fires, 2011

A fire can engulf a structure in a matter of minutes. Understanding the basic characteristics of fire and learning the proper safety practices can be the key to surviving a house or building fire.