<p> The goal of this thesis is to study the economics of computational learning. Attention is also paid to applications of computational learning models, especially Valiant's so-called `probably approximately correctly' (PAC) learning model, in econometric situations.</p><p>Specifically, an economically reasonable stopping time model of learning is the subject of two attached papers. In the rst paper, Paper A, the economics of PAC learning are considered. It is shown how a general form of the optimal stopping time bounds can be achieved using the PAC convergence rates for a `pessimistic-rational' learner in the most standard binary case of passive supervised PAC model of finite Vapnik-Chervonenkis (VC) dimension.</p><p> </p><p>The second paper, Paper B, states precisely and improves the ideas introduced in Paper A and tests them in a specific and mathematically simple case. Using the maxmin procedure of Gilboa and Schmeidler the bounds for the stopping time are expressed in terms of the largest expected error of recall, and thus, effectively, in terms of the least expected reward. The problem of locating a real number θ by testing whether x_i ≤ θ , with x_i drawn from an calculated for a range of term rates, sample costs and rewards/penalties from a recall ae included. The standard econometric situations, such as product promotion, market research, credit risk assessment, and bargaining and tenders, where such bounds could be of interest, are pointed. </p><p>These two papers are the essence of this thesis, and form it togheter with an introduction to the subject of learning.</p> / <p>Målet med denna avhandling är att studera optimering av inlärning när det finns kostnader. Speciellt studerar jag Valiants så kallade PAC-inlärningsmodell  (Probably Approximately Correctly), ofta använd inom datavetenskap. I två artiklar behandlar jag hur länge, ur ekonomisk synvinkel, inlärningsperioden bör fortsätta.</p><p>I den första artikeln visar vi hur en generell form av begränsningar av den optimala inlärningsperioden kan fås med hjälp av PAC-konvergenshastigheten för en ’pessimistiskt rationell’ studerande (i det vanligaste binära fallet av passiv PAC-inlärningsmodell med ändlig VC-dimension).</p><p>I den andra artikeln fördjupar och förbättrar vi idéerna från den första artikeln, och testar dem i en specifik situation som är matematiskt enkel. Med hjälp av Gilboa – Schmeidlers max - minprocedur  uttrycker vi begränsningarna av den optimala inlärningsperioden som funktion av det största förväntade felet och därmed som funktion av den minsta förväntade belöningen. Vi diskuterar problemet med att hitta ett reellt tal θ genom testning av huruvida x_i ≤ θ, där x_i dras från en okänd fördelning. Här tar vi också upp exempel på begränsningar av inlärningsperioden, beräknade för en mängd av diskontovärden, stickprovskostnader och belöning/straff för erinran, samt en del vanliga ekonometriska situationer där sådana begränsningar är av intresse, såsom marknadsföring av produkter, marknadsanalys, kreditriskskattning och offertförhandling.</p><p>Avhandlingen består i huvuddel av dessa två artiklar samt en kort introduktion till ekonomiska, matematiska och datavetenskapliga inlärningsmodeller.</p><p> </p>

The goal of this thesis is to study the economics of computational learning. Attention is also paid to applications of computational learning models, especially Valiant's so-called `probably approximately correctly' (PAC) learning model, in econometric situations. Specifically, an economically reasonable stopping time model of learning is the subject of two attached papers. In the rst paper, Paper A, the economics of PAC learning are considered. It is shown how a general form of the optimal stopping time bounds can be achieved using the PAC convergence rates for a `pessimistic-rational' learner in the most standard binary case of passive supervised PAC model of finite Vapnik-Chervonenkis (VC) dimension.   The second paper, Paper B, states precisely and improves the ideas introduced in Paper A and tests them in a specific and mathematically simple case. Using the maxmin procedure of Gilboa and Schmeidler the bounds for the stopping time are expressed in terms of the largest expected error of recall, and thus, effectively, in terms of the least expected reward. The problem of locating a real number θ by testing whether xi ≤ θ , with xi drawn from an calculated for a range of term rates, sample costs and rewards/penalties from a recall ae included. The standard econometric situations, such as product promotion, market research, credit risk assessment, and bargaining and tenders, where such bounds could be of interest, are pointed.  These two papers are the essence of this thesis, and form it togheter with an introduction to the subject of learning. / Målet med denna avhandling är att studera optimering av inlärning när det finns kostnader. Speciellt studerar jag Valiants så kallade PAC-inlärningsmodell  (Probably Approximately Correctly), ofta använd inom datavetenskap. I två artiklar behandlar jag hur länge, ur ekonomisk synvinkel, inlärningsperioden bör fortsätta. I den första artikeln visar vi hur en generell form av begränsningar av den optimala inlärningsperioden kan fås med hjälp av PAC-konvergenshastigheten för en ’pessimistiskt rationell’ studerande (i det vanligaste binära fallet av passiv PAC-inlärningsmodell med ändlig VC-dimension). I den andra artikeln fördjupar och förbättrar vi idéerna från den första artikeln, och testar dem i en specifik situation som är matematiskt enkel. Med hjälp av Gilboa – Schmeidlers max - minprocedur  uttrycker vi begränsningarna av den optimala inlärningsperioden som funktion av det största förväntade felet och därmed som funktion av den minsta förväntade belöningen. Vi diskuterar problemet med att hitta ett reellt tal θ genom testning av huruvida xi ≤ θ, där xi dras från en okänd fördelning. Här tar vi också upp exempel på begränsningar av inlärningsperioden, beräknade för en mängd av diskontovärden, stickprovskostnader och belöning/straff för erinran, samt en del vanliga ekonometriska situationer där sådana begränsningar är av intresse, såsom marknadsföring av produkter, marknadsanalys, kreditriskskattning och offertförhandling. Avhandlingen består i huvuddel av dessa två artiklar samt en kort introduktion till ekonomiska, matematiska och datavetenskapliga inlärningsmodeller.

ABSTRACT A continuous energy loss measurements of 63Cu, 28Si, 27AI, 24Mg, 19F, 160 and 12C ions over an energy range of(O.OG-0.65) MeV/nucleon through thin polymeric foils (Mylar, Polypropylene and Formvar) were carried out by time of flight spectrometry. The deduced experimental stopping data have been used in order to assess our proposed semi empirical formula. The proposed approach based on the Firsov and Lindhard-Scharff stopping power models is provided for well describing-the electronic stopping power of heavy ions (3 Z < 100) in various solids targets at low energy range. The '· factor, which was approximated to be z:i6 , involved in Lindhard, Scharff and Schiott (LSS) formula has been suitably modified in the light of the available experimental stopping power data. The calculated stopping power values after incorporating, effective charge z; of moving heavy ions with low velocities ( v v0z 13 ) and modified C. in LSS formula, have been found to be in close agreement with measured values in various solids targets. A reason of energy loss measurements is to obtain data that help to assess our understanding of the stopping power theories. For this, the obtained results are compared with, LSS calculations, MSTAR and SRIM-2013 predictions code.

Test-analyze-and-fix (TAAF) is the most commonly recognized method of improving system reliability. The work presented here addresses the question of when to stop testing during TAAF programs involving one-shot systems when the number of systems to be produced is predetermined and the probabilities of identifying and successfully correcting each failure mode are less than one. The goal here is to determine when to cease testing to maximize utility where utility is defined as the number of systems expected to perform successfully in the field after deployment of the lot. Two TAAF stopping rules are presented. Simulation is used to model TAAF execution under different reliability growth conditions. Four discrete reliability growth models (DRGM's) are used to generate "real world" reliability growth and to estimate reliability growth using hypothetical observed success/failure data. Ranges for the following parameters are considered: starting reliability, growth rate, maximum achievable reliability, number of systems to be produced, probability of incorrectly identifying a failure mode, and probability of an unsuccessful design modification. Conclusions are drawn regarding stopping rule performance in terms of stopping rule signal location, utility loss, achieved reliability, and fraction tested. Both rules perform well and are implementable from a practical standpoint. Specific recommendations for stopping rule implementation are given based on the controllable factors, estimation methodology and lot size. / Master of Science

Master of Science / Department of Chemical Engineering / Larry A. Glasgow / The military is interested in stopping suspicious vehicles at checkpoints or security positions while minimizing noncombatant fatalities. Preliminary work has shown that decreasing the oxygen concentration in proximity to the automobile air intake system and blocking the air flow through an automotive induction system provides the greatest probability of success for the broadest possible array of internal combustion engines. A non-lethal foam deployment system was developed that satisfies the military’s needs to stop suspicious vehicles. The foam is discharged from a pressurized tank and engulfs the air intake system of the target vehicle. The foam is drawn into the air intake and the protein additive contained in the foam would occlude pores in the air filter medium. Once the air filter was blocked, the vehicle would become immobilized so that security personnel can secure the vehicle. The work carried out in this project consisted of development and refinement of surfactant solution composition, improvement in the rate of absorption of carbon dioxide for increased foam volume, and characterization of discharge for optimum foam volume. In addition, a half-scale model apparatus was developed to test the foam’s ability to be ingested in an automotive intake system. These experiments demonstrated that the foam deployment system would stop an automobile within six seconds.