Decision-theoretic reflections on processing a fingermark

A recent publication in this journal [Neumann et al., Forensic Sci. Int. 212 (2011) 32-46] presented the results of a field study that revealed the data provided by the fingermarks not processed in a forensic science laboratory. In their study, the authors were interested in the usefulness of this additional data in order to determine whether such fingermarks would have been worth submitting to the fingermark processing workflow. Taking these ideas as a starting point, this communication here places the fingermark in its context of a case brought before a court, and examines the question of processing or not processing a fingermark from a decision-theoretic point of view. The decision-theoretic framework presented provides an answer to this question in the form of a quantified expression of the expected value of information (EVOI) associated with the processed fingermark, which can then be compared with the cost of processing the mark.

2

The quality of environmental data analysis and propagation of errors are heavily affected by the representativity of the initial sampling design [CRE 93, DEU 97, KAN 04a, LEN 06, MUL07]. Geostatistical methods such as kriging are related to field samples, whose spatial distribution is crucial for the correct detection of the phenomena. Literature about the design of environmental monitoring networks (MN) is widespread and several interesting books have recently been published [GRU 06, LEN 06, MUL 07] in order to clarify the basic principles of spatial sampling design (monitoring networks optimization) based on Support Vector Machines was proposed. Nonetheless, modelers often receive real data coming from environmental monitoring networks that suffer from problems of non-homogenity (clustering). Clustering can be related to the preferential sampling or to the impossibility of reaching certain regions.

Persistence of Transmitted HIV-1 Drug Resistance Mutations Associated with Fitness Costs and Viral Genetic Backgrounds.

Transmission of drug-resistant pathogens presents an almost-universal challenge for fighting infectious diseases. Transmitted drug resistance mutations (TDRM) can persist in the absence of drugs for considerable time. It is generally believed that differential TDRM-persistence is caused, at least partially, by variations in TDRM-fitness-costs. However, in vivo epidemiological evidence for the impact of fitness costs on TDRM-persistence is rare. Here, we studied the persistence of TDRM in HIV-1 using longitudinally-sampled nucleotide sequences from the Swiss-HIV-Cohort-Study (SHCS). All treatment-naïve individuals with TDRM at baseline were included. Persistence of TDRM was quantified via reversion rates (RR) determined with interval-censored survival models. Fitness costs of TDRM were estimated in the genetic background in which they occurred using a previously published and validated machine-learning algorithm (based on in vitro replicative capacities) and were included in the survival models as explanatory variables. In 857 sequential samples from 168 treatment-naïve patients, 17 TDRM were analyzed. RR varied substantially and ranged from 174.0/100-person-years;CI=[51.4, 588.8] (for 184V) to 2.7/100-person-years;[0.7, 10.9] (for 215D). RR increased significantly with fitness cost (increase by 1.6[1.3,2.0] per standard deviation of fitness costs). When subdividing fitness costs into the average fitness cost of a given mutation and the deviation from the average fitness cost of a mutation in a given genetic background, we found that both components were significantly associated with reversion-rates. Our results show that the substantial variations of TDRM persistence in the absence of drugs are associated with fitness-cost differences both among mutations and among different genetic backgrounds for the same mutation.

Mechanisms of action of brief alcohol interventions remain largely unknown - a narrative review.

A growing body of evidence has shown the efficacy of brief intervention (BI) for hazardous and harmful alcohol use in primary health care settings. Evidence for efficacy in other settings and effectiveness when implemented at larger scale are disappointing. Indeed, BI comprises varying content; exploring BI content and mechanisms of action may be a promising way to enhance efficacy and effectiveness. Medline and PsychInfo, as well as references of retrieved publications were searched for original research or review on active ingredients (components or mechanisms) of face-to-face BIs [and its subtypes, including brief advice and brief motivational interviewing (BMI)] for alcohol. Overall, BI active ingredients have been scarcely investigated, almost only within BMI, and mostly among patients in the emergency room, young adults, and US college students. This body of research has shown that personalized feedback may be an effective component; specific MI techniques showed mixed findings; decisional balance findings tended to suggest a potential detrimental effect; while change plan exercises, advice to reduce or stop drinking, presenting alternative change options, and moderation strategies are promising but need further study. Client change talk is a potential mediator of BMI effects; change in norm perceptions and enhanced discrepancy between current behavior and broader life goals and values have received preliminary support; readiness to change was only partially supported as a mediator; while enhanced awareness of drinking, perceived risks/benefits of alcohol use, alcohol treatment seeking, and self-efficacy were seldom studied and have as yet found no significant support as such. Research is obviously limited and has provided no clear and consistent evidence on the mechanisms of alcohol BI. How BI achieves the effects seen in randomized trials remains mostly unknown and should be investigated to inform the development of more effective interventions.

Functional error modeling for uncertainty quantification in hydrogeology

Approximate models (proxies) can be employed to reduce the computational costs of estimating uncertainty. The price to pay is that the approximations introduced by the proxy model can lead to a biased estimation. To avoid this problem and ensure a reliable uncertainty quantification, we propose to combine functional data analysis and machine learning to build error models that allow us to obtain an accurate prediction of the exact response without solving the exact model for all realizations. We build the relationship between proxy and exact model on a learning set of geostatistical realizations for which both exact and approximate solvers are run. Functional principal components analysis (FPCA) is used to investigate the variability in the two sets of curves and reduce the dimensionality of the problem while maximizing the retained information. Once obtained, the error model can be used to predict the exact response of any realization on the basis of the sole proxy response. This methodology is purpose-oriented as the error model is constructed directly for the quantity of interest, rather than for the state of the system. Also, the dimensionality reduction performed by FPCA allows a diagnostic of the quality of the error model to assess the informativeness of the learning set and the fidelity of the proxy to the exact model. The possibility of obtaining a prediction of the exact response for any newly generated realization suggests that the methodology can be effectively used beyond the context of uncertainty quantification, in particular for Bayesian inference and optimization.

Transferring control demands across incidental learning tasks - Stronger sequence usage in serial reaction task after shortcut option in letter string checking

After incidentally learning about a hidden regularity, participants can either continue to solve the task as instructed or, alternatively, apply a shortcut. Past research suggests that the amount of conflict implied by adopting a shortcut seems to bias the decision for vs. against continuing instruction-coherent task processing. We explored whether this decision might transfer from one incidental learning task to the next. Theories that conceptualize strategy change in incidental learning as a learning-plus-decision phenomenon suggest that high demands to adhere to instruction-coherent task processing in Task 1 will impede shortcut usage in Task 2, whereas low control demands will foster it. We sequentially applied two established incidental learning tasks differing in stimuli, responses and hidden regularity (the alphabet verification task followed by the serial reaction task, SRT). While some participants experienced a complete redundancy in the task material of the alphabet verification task (low demands to adhere to instructions), for others the redundancy was only partial. Thus, shortcut application would have led to errors (high demands to follow instructions). The low control demand condition showed the strongest usage of the fixed and repeating sequence of responses in the SRT. The transfer results are in line with the learning-plus-decision view of strategy change in incidental learning, rather than with resource theories of self-control.

Päätöspuun käyttö tutkamaalien initialisoinnissa

Suomen ilmatilaa valvotaan reaaliaikaisesti, pääasiassa ilmavalvontatutkilla. Ilmatilassa on lentokoneiden lisäksi paljon muitakin kohteita, jotka tutka havaitsee. Tutka lähettää nämä tiedot edelleen ilmavalvontajärjestelmään. Ilmavalvontajärjestelmä käsittelee tiedot, sekä lähettää ne edelleen esitysjärjestelmään. Esitysjärjestelmässä tiedot esitetään synteettisinä merkkeinä, seurantoina joista käytetään nimitystä träkki. Näiden tietojen puitteissa sekä oman ammattitaitonsa perusteella ihmiset tekevät päätöksiä. Tämän työn tarkoituksena on tutkia tutkan havaintoja träkkien initialisointipisteessä siten, että voitaisiin määritellä tyypillinen rakenne sille mikä on oikea ja mikä väärä tai huono träkki. Tämän lisäksi tulisi ennustaa, mitkä Irakeista eivät aiheudu ilma- aluksista. Saadut tulokset voivat helpottaa työtä havaintojen tulkinnassa - jokainen lintuparvi ei ole ehdokas seurannaksi. Havaintojen luokittelu voidaan tehdä joko neurolaskennalla tai päätöspuulla. Neurolaskenta tehdään neuroverkoilla, jotka koostuvat neuroneista. Päätöspuu- luokittelijat ovat oppivia tietorakenteita kuten neuroverkotkin. Yleisin päätöpuu on binääripuu. Tämän työn tavoitteena on opettaa päätöspuuluokittelija havaintojen avulla siten, että se pystyy luokittelemaan väärät havainnot oikeista. Neurolaskennan mahdollisuuksia tässä työssä ei käsitellä kuin teoreettisesti. Työn tuloksena voi todeta, että päätöspuuluokittelijat ovat erittäin kykeneviä erottamaan oikeat havainnot vääristä. Vaikka tulokset olivat rohkaiseva, lisää tutkimusta tarvitaan määrittelemään luotettavammin tekijät, jotka parhaiten suorittavat luokittelun.

Collective decision making in a heterogeneous environment: Lasius niger colonies preferentially forage at easy to learn locations

Many ants forage in complex environments and use a combination of trail pheromone information and route memory to navigate between food sources and the nest. Previous research has shown that foraging routes differ in how easily they are learned. In particular, it is easier to learn feeding locations that are reached by repeating (e.g. left-left or right-right) than alternating choices (left-right or right-left) along a route with two T-bifurcations. This raises the hypothesis that the learnability of the feeding sites may influence overall colony foraging patterns. We studied this in the mass-recruiting ant Lasius niger. We used mazes with two T-bifurcations, and allowed colonies to exploit two equidistant food sources that differed in how easily their locations were learned. In experiment 1, learnability was manipulated by using repeating versus alternating routes from nest to feeder. In experiment 2, we added visual landmarks along the route to one food source. Our results suggest that colonies preferentially exploited the feeding site that was easier to learn. This was the case even if the more difficult to learn feeding site was discovered first. Furthermore, we show that these preferences were at least partly caused by lower error rates (experiment 1) and greater foraging speeds (experiment 2) of foragers visiting the more easily learned feeder locations. Our results indicate that the learnability of feeding sites is an important factor influencing collective foraging patterns of ant colonies under more natural conditions, given that in natural environments foragers often face multiple bifurcations on their way to food sources.

Error models in hydrogeology applications

Notre consommation en eau souterraine, en particulier comme eau potable ou pour l'irrigation, a considérablement augmenté au cours des années. De nombreux problèmes font alors leur apparition, allant de la prospection de nouvelles ressources à la remédiation des aquifères pollués. Indépendamment du problème hydrogéologique considéré, le principal défi reste la caractérisation des propriétés du sous-sol. Une approche stochastique est alors nécessaire afin de représenter cette incertitude en considérant de multiples scénarios géologiques et en générant un grand nombre de réalisations géostatistiques. Nous rencontrons alors la principale limitation de ces approches qui est le coût de calcul dû à la simulation des processus d'écoulements complexes pour chacune de ces réalisations. Dans la première partie de la thèse, ce problème est investigué dans le contexte de propagation de l'incertitude, oú un ensemble de réalisations est identifié comme représentant les propriétés du sous-sol. Afin de propager cette incertitude à la quantité d'intérêt tout en limitant le coût de calcul, les méthodes actuelles font appel à des modèles d'écoulement approximés. Cela permet l'identification d'un sous-ensemble de réalisations représentant la variabilité de l'ensemble initial. Le modèle complexe d'écoulement est alors évalué uniquement pour ce sousensemble, et, sur la base de ces réponses complexes, l'inférence est faite. Notre objectif est d'améliorer la performance de cette approche en utilisant toute l'information à disposition. Pour cela, le sous-ensemble de réponses approximées et exactes est utilisé afin de construire un modèle d'erreur, qui sert ensuite à corriger le reste des réponses approximées et prédire la réponse du modèle complexe. Cette méthode permet de maximiser l'utilisation de l'information à disposition sans augmentation perceptible du temps de calcul. La propagation de l'incertitude est alors plus précise et plus robuste. La stratégie explorée dans le premier chapitre consiste à apprendre d'un sous-ensemble de réalisations la relation entre les modèles d'écoulement approximé et complexe. Dans la seconde partie de la thèse, cette méthodologie est formalisée mathématiquement en introduisant un modèle de régression entre les réponses fonctionnelles. Comme ce problème est mal posé, il est nécessaire d'en réduire la dimensionnalité. Dans cette optique, l'innovation du travail présenté provient de l'utilisation de l'analyse en composantes principales fonctionnelles (ACPF), qui non seulement effectue la réduction de dimensionnalités tout en maximisant l'information retenue, mais permet aussi de diagnostiquer la qualité du modèle d'erreur dans cet espace fonctionnel. La méthodologie proposée est appliquée à un problème de pollution par une phase liquide nonaqueuse et les résultats obtenus montrent que le modèle d'erreur permet une forte réduction du temps de calcul tout en estimant correctement l'incertitude. De plus, pour chaque réponse approximée, une prédiction de la réponse complexe est fournie par le modèle d'erreur. Le concept de modèle d'erreur fonctionnel est donc pertinent pour la propagation de l'incertitude, mais aussi pour les problèmes d'inférence bayésienne. Les méthodes de Monte Carlo par chaîne de Markov (MCMC) sont les algorithmes les plus communément utilisés afin de générer des réalisations géostatistiques en accord avec les observations. Cependant, ces méthodes souffrent d'un taux d'acceptation très bas pour les problèmes de grande dimensionnalité, résultant en un grand nombre de simulations d'écoulement gaspillées. Une approche en deux temps, le "MCMC en deux étapes", a été introduite afin d'éviter les simulations du modèle complexe inutiles par une évaluation préliminaire de la réalisation. Dans la troisième partie de la thèse, le modèle d'écoulement approximé couplé à un modèle d'erreur sert d'évaluation préliminaire pour le "MCMC en deux étapes". Nous démontrons une augmentation du taux d'acceptation par un facteur de 1.5 à 3 en comparaison avec une implémentation classique de MCMC. Une question reste sans réponse : comment choisir la taille de l'ensemble d'entrainement et comment identifier les réalisations permettant d'optimiser la construction du modèle d'erreur. Cela requiert une stratégie itérative afin que, à chaque nouvelle simulation d'écoulement, le modèle d'erreur soit amélioré en incorporant les nouvelles informations. Ceci est développé dans la quatrième partie de la thèse, oú cette méthodologie est appliquée à un problème d'intrusion saline dans un aquifère côtier. -- Our consumption of groundwater, in particular as drinking water and for irrigation, has considerably increased over the years and groundwater is becoming an increasingly scarce and endangered resource. Nofadays, we are facing many problems ranging from water prospection to sustainable management and remediation of polluted aquifers. Independently of the hydrogeological problem, the main challenge remains dealing with the incomplete knofledge of the underground properties. Stochastic approaches have been developed to represent this uncertainty by considering multiple geological scenarios and generating a large number of realizations. The main limitation of this approach is the computational cost associated with performing complex of simulations in each realization. In the first part of the thesis, we explore this issue in the context of uncertainty propagation, where an ensemble of geostatistical realizations is identified as representative of the subsurface uncertainty. To propagate this lack of knofledge to the quantity of interest (e.g., the concentration of pollutant in extracted water), it is necessary to evaluate the of response of each realization. Due to computational constraints, state-of-the-art methods make use of approximate of simulation, to identify a subset of realizations that represents the variability of the ensemble. The complex and computationally heavy of model is then run for this subset based on which inference is made. Our objective is to increase the performance of this approach by using all of the available information and not solely the subset of exact responses. Two error models are proposed to correct the approximate responses follofing a machine learning approach. For the subset identified by a classical approach (here the distance kernel method) both the approximate and the exact responses are knofn. This information is used to construct an error model and correct the ensemble of approximate responses to predict the "expected" responses of the exact model. The proposed methodology makes use of all the available information without perceptible additional computational costs and leads to an increase in accuracy and robustness of the uncertainty propagation. The strategy explored in the first chapter consists in learning from a subset of realizations the relationship between proxy and exact curves. In the second part of this thesis, the strategy is formalized in a rigorous mathematical framework by defining a regression model between functions. As this problem is ill-posed, it is necessary to reduce its dimensionality. The novelty of the work comes from the use of functional principal component analysis (FPCA), which not only performs the dimensionality reduction while maximizing the retained information, but also allofs a diagnostic of the quality of the error model in the functional space. The proposed methodology is applied to a pollution problem by a non-aqueous phase-liquid. The error model allofs a strong reduction of the computational cost while providing a good estimate of the uncertainty. The individual correction of the proxy response by the error model leads to an excellent prediction of the exact response, opening the door to many applications. The concept of functional error model is useful not only in the context of uncertainty propagation, but also, and maybe even more so, to perform Bayesian inference. Monte Carlo Markov Chain (MCMC) algorithms are the most common choice to ensure that the generated realizations are sampled in accordance with the observations. Hofever, this approach suffers from lof acceptance rate in high dimensional problems, resulting in a large number of wasted of simulations. This led to the introduction of two-stage MCMC, where the computational cost is decreased by avoiding unnecessary simulation of the exact of thanks to a preliminary evaluation of the proposal. In the third part of the thesis, a proxy is coupled to an error model to provide an approximate response for the two-stage MCMC set-up. We demonstrate an increase in acceptance rate by a factor three with respect to one-stage MCMC results. An open question remains: hof do we choose the size of the learning set and identify the realizations to optimize the construction of the error model. This requires devising an iterative strategy to construct the error model, such that, as new of simulations are performed, the error model is iteratively improved by incorporating the new information. This is discussed in the fourth part of the thesis, in which we apply this methodology to a problem of saline intrusion in a coastal aquifer.

New kernel functions and learning methods for text and data mining

Recent advances in machine learning methods enable increasingly the automatic construction of various types of computer assisted methods that have been difficult or laborious to program by human experts. The tasks for which this kind of tools are needed arise in many areas, here especially in the fields of bioinformatics and natural language processing. The machine learning methods may not work satisfactorily if they are not appropriately tailored to the task in question. However, their learning performance can often be improved by taking advantage of deeper insight of the application domain or the learning problem at hand. This thesis considers developing kernel-based learning algorithms incorporating this kind of prior knowledge of the task in question in an advantageous way. Moreover, computationally efficient algorithms for training the learning machines for specific tasks are presented. In the context of kernel-based learning methods, the incorporation of prior knowledge is often done by designing appropriate kernel functions. Another well-known way is to develop cost functions that fit to the task under consideration. For disambiguation tasks in natural language, we develop kernel functions that take account of the positional information and the mutual similarities of words. It is shown that the use of this information significantly improves the disambiguation performance of the learning machine. Further, we design a new cost function that is better suitable for the task of information retrieval and for more general ranking problems than the cost functions designed for regression and classification. We also consider other applications of the kernel-based learning algorithms such as text categorization, and pattern recognition in differential display. We develop computationally efficient algorithms for training the considered learning machines with the proposed kernel functions. We also design a fast cross-validation algorithm for regularized least-squares type of learning algorithm. Further, an efficient version of the regularized least-squares algorithm that can be used together with the new cost function for preference learning and ranking tasks is proposed. In summary, we demonstrate that the incorporation of prior knowledge is possible and beneficial, and novel advanced kernels and cost functions can be used in algorithms efficiently.

To analyse a trace or not? Evaluating the decision-making process in the criminal investigation

In order to broaden our knowledge and understanding of the decision steps in the criminal investigation process, we started by evaluating the decision to analyse a trace and the factors involved in this decision step. This decision step is embedded in the complete criminal investigation process, involving multiple decision and triaging steps. Considering robbery cases occurring in a geographic region during a 2-year-period, we have studied the factors influencing the decision to submit biological traces, directly sampled on the scene of the robbery or on collected objects, for analysis. The factors were categorised into five knowledge dimensions: strategic, immediate, physical, criminal and utility and decision tree analysis was carried out. Factors in each category played a role in the decision to analyse a biological trace. Interestingly, factors involving information available prior to the analysis are of importance, such as the fact that a positive result (a profile suitable for comparison) is already available in the case, or that a suspect has been identified through traditional police work before analysis. One factor that was taken into account, but was not significant, is the matrix of the trace. Hence, the decision to analyse a trace is not influenced by this variable. The decision to analyse a trace first is very complex and many of the tested variables were taken into account. The decisions are often made on a case-by-case basis.

Design and evaluation of learning algorithms for dynamic resource management in virtual networks

Network virtualisation is considerably gaining attentionas a solution to ossification of the Internet. However, thesuccess of network virtualisation will depend in part on how efficientlythe virtual networks utilise substrate network resources.In this paper, we propose a machine learning-based approachto virtual network resource management. We propose to modelthe substrate network as a decentralised system and introducea learning algorithm in each substrate node and substrate link,providing self-organization capabilities. We propose a multiagentlearning algorithm that carries out the substrate network resourcemanagement in a coordinated and decentralised way. The taskof these agents is to use evaluative feedback to learn an optimalpolicy so as to dynamically allocate network resources to virtualnodes and links. The agents ensure that while the virtual networkshave the resources they need at any given time, only the requiredresources are reserved for this purpose. Simulations show thatour dynamic approach significantly improves the virtual networkacceptance ratio and the maximum number of accepted virtualnetwork requests at any time while ensuring that virtual networkquality of service requirements such as packet drop rate andvirtual link delay are not affected.

Learning Preferences with Kernel-Based Methods

Learning of preference relations has recently received significant attention in machine learning community. It is closely related to the classification and regression analysis and can be reduced to these tasks. However, preference learning involves prediction of ordering of the data points rather than prediction of a single numerical value as in case of regression or a class label as in case of classification. Therefore, studying preference relations within a separate framework facilitates not only better theoretical understanding of the problem, but also motivates development of the efficient algorithms for the task. Preference learning has many applications in domains such as information retrieval, bioinformatics, natural language processing, etc. For example, algorithms that learn to rank are frequently used in search engines for ordering documents retrieved by the query. Preference learning methods have been also applied to collaborative filtering problems for predicting individual customer choices from the vast amount of user generated feedback. In this thesis we propose several algorithms for learning preference relations. These algorithms stem from well founded and robust class of regularized least-squares methods and have many attractive computational properties. In order to improve the performance of our methods, we introduce several non-linear kernel functions. Thus, contribution of this thesis is twofold: kernel functions for structured data that are used to take advantage of various non-vectorial data representations and the preference learning algorithms that are suitable for different tasks, namely efficient learning of preference relations, learning with large amount of training data, and semi-supervised preference learning. Proposed kernel-based algorithms and kernels are applied to the parse ranking task in natural language processing, document ranking in information retrieval, and remote homology detection in bioinformatics domain. Training of kernel-based ranking algorithms can be infeasible when the size of the training set is large. This problem is addressed by proposing a preference learning algorithm whose computation complexity scales linearly with the number of training data points. We also introduce sparse approximation of the algorithm that can be efficiently trained with large amount of data. For situations when small amount of labeled data but a large amount of unlabeled data is available, we propose a co-regularized preference learning algorithm. To conclude, the methods presented in this thesis address not only the problem of the efficient training of the algorithms but also fast regularization parameter selection, multiple output prediction, and cross-validation. Furthermore, proposed algorithms lead to notably better performance in many preference learning tasks considered.

Kernel-Based Ranking. Methods for Learning and Performance Estimation

Machine learning provides tools for automated construction of predictive models in data intensive areas of engineering and science. The family of regularized kernel methods have in the recent years become one of the mainstream approaches to machine learning, due to a number of advantages the methods share. The approach provides theoretically well-founded solutions to the problems of under- and overfitting, allows learning from structured data, and has been empirically demonstrated to yield high predictive performance on a wide range of application domains. Historically, the problems of classification and regression have gained the majority of attention in the field. In this thesis we focus on another type of learning problem, that of learning to rank. In learning to rank, the aim is from a set of past observations to learn a ranking function that can order new objects according to how well they match some underlying criterion of goodness. As an important special case of the setting, we can recover the bipartite ranking problem, corresponding to maximizing the area under the ROC curve (AUC) in binary classification. Ranking applications appear in a large variety of settings, examples encountered in this thesis include document retrieval in web search, recommender systems, information extraction and automated parsing of natural language. We consider the pairwise approach to learning to rank, where ranking models are learned by minimizing the expected probability of ranking any two randomly drawn test examples incorrectly. The development of computationally efficient kernel methods, based on this approach, has in the past proven to be challenging. Moreover, it is not clear what techniques for estimating the predictive performance of learned models are the most reliable in the ranking setting, and how the techniques can be implemented efficiently. The contributions of this thesis are as follows. First, we develop RankRLS, a computationally efficient kernel method for learning to rank, that is based on minimizing a regularized pairwise least-squares loss. In addition to training methods, we introduce a variety of algorithms for tasks such as model selection, multi-output learning, and cross-validation, based on computational shortcuts from matrix algebra. Second, we improve the fastest known training method for the linear version of the RankSVM algorithm, which is one of the most well established methods for learning to rank. Third, we study the combination of the empirical kernel map and reduced set approximation, which allows the large-scale training of kernel machines using linear solvers, and propose computationally efficient solutions to cross-validation when using the approach. Next, we explore the problem of reliable cross-validation when using AUC as a performance criterion, through an extensive simulation study. We demonstrate that the proposed leave-pair-out cross-validation approach leads to more reliable performance estimation than commonly used alternative approaches. Finally, we present a case study on applying machine learning to information extraction from biomedical literature, which combines several of the approaches considered in the thesis. The thesis is divided into two parts. Part I provides the background for the research work and summarizes the most central results, Part II consists of the five original research articles that are the main contribution of this thesis.