Search

Microarray based transcription profiling is now a consolidated methodology and has widespread use in areas such as pharmacogenomics, diagnostics and drug target identification. Large-scale microarray studies are also becoming crucial to a new way of conceiving experimental biology. A main issue in microarray transcription profiling is data analysis and mining. When microarrays became a methodology of general use, considerable effort was made to produce algorithms and methods for the identification of differentially expressed genes. More recently, the focus has switched to algorithms and database development for microarray data mining. Furthermore, the evolution of microarray technology is allowing researchers to grasp the regulative nature of transcription, integrating basic expression analysis with mRNA characteristics, i.e. exon-based arrays, and with DNA characteristics, i.e. comparative genomic hybridization, single nucleotide polymorphism, tiling and promoter structure. In this article, we will review approaches used to detect differentially expressed genes and to link differential expression to specific biological functions.

In recent years, the KDD process has been advocated to be an iterative and interactive process. It is seldom the case that a user is able to answer immediately all his questions on date with a single query. On the contrary, the work-flow of the typical user consists of several steps in which he/she iteratively refines the extracted knowledge by inspecting previous results and posing new queries. Given this view of the KDD process, in order to reduce the computational effort, it becomes crucial to have KDD systems that are able to exploit past results. This is especially true in environments in which the system knowledge base is the result of many discoveries on data made separately by the collaborative effort of different users. In this paper, we consider the problem of mining frequent association rules from database relations. We first model a general, constraint-based, mining language for this task. Then, we propose an algorithm that answers such queries reusing past results. In particular, this solution is effective for a new class of constraints, called context dependent, which are more difficult than the traditionally studied item dependent constraints. Nevertheless, we show that some typical queries of important application domains, such as market stock trading, analysis of web log, and gene microarrays in bioinformatics, have context-dependent constraints. We show with a set of experiments in these application domains that the proposed solution with an incremental approach is both effective and viable.

The proliferation of Infrared technology and imaging systems enables a different perspective to tackle many computer vision problems in defense and security applications. Infrared images are widely used by the law enforcement, Homeland Security and military organizations to achieve a significant advantage or situational awareness, and thus is vital to protect these data against malicious attacks. Concurrently, sophisticated malware are developed which are able to disrupt the security and integrity of these digital media. For instance, illegal distribution and manipulation are possible malicious attacks to the digital objects. In this paper we explore the use of a new layer of defense for the integrity of the infrared images through the aid of information hiding techniques such as watermarking. In this context, we analyze the efficiency of several optimal decoding schemes for the watermark inserted into the Singular Value Decomposition (SVD) domain of the IR images using an additive spread spectrum (SS) embedding framework. In order to use the singular values (SVs) of the IR images with the SS embedding we adopt several restrictions that ensure that the values of the SVs will maintain their statistics. For both the optimal maximum likelihood decoder and sub-optimal decoders we assume that the PDF of SVs can be modeled by the Weibull distribution. Furthermore, we investigate the challenges involved in protecting and assuring the integrity of IR images such as data complexity and the error probability behavior, i.e., the probability of detection and the probability of false detection, for the applied optimal decoders. By taking into account the efficiency and the necessary auxiliary information for decoding the watermark, we discuss the suitable decoder for various operating situations. Experimental results are carried out on a large dataset of IR images to show the imperceptibility and efficiency of the proposed scheme against various attack scenarios.

This paper describes a module that extracts rules or frequent patterns through data mining from a large database fed by targets detected by a Mission System installed on an unmanned airborne platform and the associated ground station to discover anomalies in local traffic. It has been demonstrated that the module is able to detect all tracks or targets present in the ground truth and also the paths followed by each tracks. Traffic anomalies can be detected by observing differences in extracted rules in reference missions compared to the current mission. The module will significantly reduce the operator workload as it can operate autonomously‥

In this paper we show that a recently proposed fragile watermarking scheme by Rawat et al. does not detect and localize tampering, therefore cannot be used for authentication applications. The problem lies in that the scheme embeds an authentication code into the LSBs of pixels without taking into consideration the image content. To overcome this issue the authentication should be combined with the first seven bits of the image pixels, and in this paper a revision in this sense is proposed.

This paper presents a security analysis of robust watermarking methods based on Schur decomposition in a general scenario. The security is defined as the difficulty to remove the watermark and to estimate the secrets used in the embedding process, supposing that the adversary possesses several watermarked digital contents. The theoretical analysis and extensive experimental results carried out prove that these schemes fail to secure the digital contents against malicious attacks.

In recent years, the necessity to protect multimedia content from illegal copying has been made more critical by the advent of digital technology. Singular Value Decomposition (SVD) received much attention from the watermarking community due to its appealing properties, such as the invariance of singular values (SVs) to common image processing operations and geometric transforms, like rotation, translation and scaling. However, most of watermarking schemes require a lot of prior information to detect the watermark and in addition they are not secure. In this paper we investigate the efficiency of optimal decoding schemes for the watermark inserted into the SVD domain of the host images using an additive spread spectrum (SS) embedding framework. In order to use the singular values (SVs) with the SS embedding we adopt several restrictions that ensure that the values of the SVs remain real, positive and sorted. For both the optimal maximum likelihood decoder and sub-optimal decoders we assume that the PDF of SVs can be modeled by the Weibull distribution. Furthermore, we investigate the error probability behavior, i.e., the probability of detection and the probability of false detection, for the applied optimal decoders. By taking into account the efficiency and the necessary auxiliary information for decoding the watermark, we discuss the suitable decoder for various operating situations.

This paper proposes a method for refining numerical constants occurring in rules of a knowledge base expressed in a first order logic language. The method consists in tuning numerical parameters by performing error gradient descent. The knowledge base to be refined can be manually handcrafted or automatically acquired by a symbolic relational learner, able to deal with numerical features. The results of an experimental analysis performed on four case studies show that the refinement step can be effective in improving classification performances.

This paper describes a representation framework that offers a unifying platform for alternative systems, which learn concepts in First Order Logics. The main aspects of this framework are discussed. First of all, the separation between the hypothesis logical language (a version of the VL21 language) and the representation of data by means of a relational database is motivated. Then, the functional layer between data and hypotheses, which makes the data accessible by the logical level through a set of abstract properties is described. A novelty, in the hypothesis representation language, is the introduction of the construct of internal disjunction; such a construct, first used by the AQ and Induce systems, is here made operational via a set of algorithms, capable to learn it, for both the discrete and the continuous-valued attributes case. These algorithms are embedded in learning systems (SMART+, REGAL, SNAP, WHY, RTL) using different paradigms (symbolic, genetic or connectionist), thus realizing an effective integration among them; in fact, categorical and numerical attributes can be handled in a uniform way. In order to exemplify the effectiveness of the representation framework and of the multistrategy integration, the results obtained by the above systems in some application domains are summarized.

During the last decade, the evolution of the competition law enforcement has reinvigorated the debate concerning the compatibility of the EU enforcement regime with fundamental rights. In particular, two factors have increased the relevance of the protection of fundamental rights in the context of competition law enforcement: the entry into force of the Lisbon Treaty (incorporating the EU Charter of Fundamental Rights) and the progressive criminalization of competition law enforcement. One of the tensions between competition law enforcement and the protection of fundamental rights concerns the right to a fair trial included in Article 6 of the European Convention of Human Rights (ECHR), as well as in Article 47(2) EU Charter. Due to the technical nature of competition law analysis, the EU and national courts have traditionally granted a certain margin of discretion to the competition authorities. In the aftermath of the Menarini judgment of the European Court of Human Rights (ECtHR), the question is whether such margin of discretion still complies with the requirement of full judicial review introduced in Menarini. This chapter reviews the ongoing debate on compatibility of EU competition law enforcement with the fundamental rights enshrined in the EU Charter (Sect.1), discusses the requirement of full judicial review introduced by the ECtHR in Menarini (Sect. 2), and analyzes the compatibility of the requirement of full judicial review with the standard applied by the Court of Justice of European Union (CJEU) in antitrust cases (Sect. 3). While in its recent jurisprudence the CJEU confirms the compatibility of its standard of judicial review with Article 47 of the EU Charter, it remains unclear whether the ECtHR would achieve the same conclusion in relation to Article 6 ECHR. The issue of compatibility of judicial review with the requirements of fair trial remains open at the national level where the courts often continue to limit their review to the procedural aspects of the competition authority’s decisions.

Since the Treaty of Rome, State aid rules have been enforced at a supranational level in the EU. However, in the framework of the past EU enlargement to the countries of Central and Eastern Europe (CEECs), the European Commission has tried to decentralize the enforcement of State aid rules by requiring CEECs to establish national State aid monitoring authorities. Similarly, during the recent years the European Commission has encouraged countries of South-East Europe.

In this paper we present a simple and elegant technique for fragile image watermarking inthe Karhunen-Loeve transform (KLT) domain with the objective of content integrity. The proposed method inserts a binary watermark into some KLT coefficients defined using a secret key image. The coefficients are modified according to a rule based on a weighted-modulo sum. The KLT is applied to contiguous blocks of the host image and a Genetic Algorithm (GA) is used to modify the pixel values in such a way that the resulting blocks contain the watermark. Given that the KLT space of insertion of the watermark is kept secret, the security of the method is strong. We experimentally demonstrate that the algorithm achieves an excellent sensitivity even to small modifications of the watermarked image.

There is accumulating evidence that driver distraction is a leading cause of vehicle crashes and incidents. In particular, increased use of so-called in-vehicle information systems (IVIS) and partially autonomous driving assistance systems (PADAS) have raised important and growing safety concerns. Thus, detecting the driver's state is of paramount importance, to adapt IVIS and PADAS accordingly, therefore avoiding or mitigating their possible negative effects. The purpose of this paper is to show a method for the nonintrusive and real-time detection of visual distraction, using vehicle dynamics data and without using the eye-tracker data as inputs to classifiers. Specifically, we present and compare different models that are based on well-known machine learning (ML) methods. Data for training the models were collected using a static driving simulator, with real human subjects performing a specific secondary task [i.e., a surrogate visual research task (SURT)] while driving. Different training methods, model characteristics, and feature selection criteria have been compared. Based on our results, using a support vector machine (SVM) has outperformed all the other ML methods, providing the highest classification rate for most of the subjects. Potential applications of this paper include the design of an adaptive IVIS and of a “smarter” PADAS.

Summary: RRE allows the extraction of non-coding regions surrounding a coding sequence [i.e. gene upstream region, 5′-untranslated region (5′-UTR), introns, 3′-UTR, downstream region] from annotated genomic datasets available at NCBI. Availability: RRE parser and web-based interface are accessible at http://www.bioinformatica.unito.it/bioinformatics/rre/rre.html

[Figure][1] [![Graphic][3] ][3][![Graphic][4] ][4] A 57-year-old woman without coronary risk factors presented with non–ST-segment elevation myocardial infarction. Angiography revealed a normal anterior descending artery with lumen irregularities in the middle part (A, B, C

Using a sample of non-financial listed firms located in the Euro area, I investigate the determinants of capital structure choices. In line with the traditional theoretical approach, and in contrast with previous empirical literature, I use a full market-value measure of leverage, estimated with the Black-Scholes-Merton model for pricing corporate debt. Differences in national tax rules are explicitly taken into account when estimating firm-specific marginal tax rates. In the cross-section regressions for firm leverage I find that some variables have similar effects across countries, while others play a different role. Assets risk, measured as the annualized volatility of the market enterprise value, is the best predictor of observed leverage ratios. The sector of activity is a significant factor in determining leverage decisions, even when assets’ risk is taken into account. Despite the monetary union and the increased financial integration in the Euro Area, nationality is still a significant determinant of capital structure decisions.

This Working Paper is a collection of contributions presented during the Workshop Framework for Economic Development in EU External Relations on 22nd and 23rd January 2010. The contributions explore various facets of what is known as economic development and what with time has incorporated many new aspects and developed links with other policy fields. Starting off with the traditional realm of the development policies – trade – the paper shifts to explore how the solely economic notion of development has been enhanced through the introduction of the concept of sustainable development, and how this concept was integrated into two policy areas – that of migration and human rights. Each of the contributors provides an in-depth analysis of a respective facet which build up a very complex picture of what economic development in EU external relations has become.

In this paper a fragile watermarking technique based on Karhunen-Loeve transform (KLT) and genetic algorithms (GA) is proposed. To achieve high sensibility to content manipulations, the proposed algorithm slightly modifies the middle-frequency KLT coefficients while maintaining the visual imperceptibility of the watermark (i.e., the average peak signal-to-noise ratio reported was 52.49 dB). It uses a GA to compensate the errors in the extraction of the authentication information, which is a pseudo-random generated watermark sequence. The experimental tests carried out show that the scheme is able to detect pixel-level alterations localizing the block in which the pixel(s) was modified.

Discovery biological organisation of the cell in modules network is a challenging task. Currently, approaches based on a controlled vocabulary, as Gene Ontology, to identify the function similarity among a pair of genes have developed. We present RGO: a rewriting of the GO aiming at obtaining a more compact and informative ontology, leading to closer biological regulated GO terms. The RGO will help researcher to easily identify the genes belonging to the same network module without the need of additional data.

A prime goal in systems biology is the comprehensive use of existing high-throughput genomic datasets to gain a better understanding of chromatin organization and genome function. In this report, we use chromatin immunoprecipitation (ChIP) data that map protein-binding sites on the genome, and Hi-C data that map interactions between DNA fragments in the genome in an integrative approach. We first reanalyzed the contact map of the human genome as determined with Hi-C and found that long-range interactions are highly nonrandom; the same DNA fragments are often found interacting together. We then show using ChIP data that these interactions can be explained by the action of the CCCTC-binding factor (CTCF). These CTCF-mediated interactions are found both within chromosomes and in between different chromosomes. This makes CTCF a major organizer of both the structure of the chromosomal fiber within each individual chromosome and of the chromosome territories within the cell nucleus.

The problem of learning a classifier from examples is a fundamental task in Machine Learnig and is nowadays actively studied. Most approaches follow a Discrimination based paradigm, where the aim is to find the best way to separate the examples of one class from those of other classes. A less popular approach follows a Characterization based paradigm, where the aim is to build a description (model) of every class from examples, and use this model to recognizes instances of that class. This paper focuses on the specific task of classifying and tagging symbolic sequences, by introducing a Characterization approach, based on Structured Hidden Markov Models, and compares its perfomances against a widely used discriminative approach, i.e. kernel machines. This task is particulary relevant to several applications in many fields, such as molecular biology, web log analysis, network traffic analysis, user profiling, and so on. In order to assess the validity of the proposed approach, an artificial benchmark has been designed in such a way that the regularities to be discovered are well known and allow for a controlled evaluation of the real capabilities the learning algorithms investigated. The obtained results allow to point out the major advantages and weaknesses of the investigated approaches in the specific classification task addressed.

The paper analyses the challenges faced by the National Competition Authorities (NCA) of the emerging economies in the application of the econometric tests usually relied on in the merger control analysis. The relevant market and the degree of market concentration are usually defined on the basis of two econometric tests, namely, the Small but Significant and Non-Transitory Price Increase (SSNIP) test and the Herfindahl-Hirschman Index (HHI). Following the example of the US and the EU Merger Guidelines, the newly established NCAs of the emerging economies often refer to these two econometric tests in their merger guidelines without taking in consideration the challenges that they will have to face in applying them. The paper is a case study on Brazil and Argentina, and it aims at defining such challenges. The paper argues in favour of granting a broader margin of discretion to the NCAs of the emerging economies in applying such tests and in favour of the possibility to introduce alternative parameters to apply such tests in comparison to those ones identified by the US and the EU Merger Guidelines.

The multiple sequence alignment problem is one the most common task in the analysis of sequential data, especially in bioinformatics. In this paper, we propose to use a genetic algorithm to compute a multiple sequence alignment, by optimizing a simple scoring function. Even though the idea of using genetic algorithms is not new, the presented approach differs in the representation of the multiple alignment and in the simplicity of the genetic operators. The results so far obtained are reported and discussed in this paper.

This paper presents a new approach to learning diagnostic knowledge, based on the use of a causal model of the domain and abductive reasoning. Justifications supplied by the casual model serve both to focus the search during one-step learning and to localize failures and propose changes during knowledge refinement.

The huge volume of gene expression data produced by microarrays and other high-throughput techniques has encouraged the development of new computational techniques to evaluate the data and to formulate new biological hypotheses. To this purpose, co-clustering techniques are widely used: these identify groups of genes that show similar activity patterns under a specific subset of the experimental conditions by measuring the similarity in expression within these groups. However, in many applications, distance metrics based only on expression levels fail in capturing biologically meaningful clusters. We propose a methodology in which a standard expression-based co-clustering algorithm is enhanced by sets of constraints which take into account the similarity/dissimilarity (inferred by the Gene Ontology, GO) between pairs of genes. Our approach minimizes the intervention of the analyst within the co-clustering process. It provides meaningful co-clusters whose discovery and interpretation is increased by embedding GO annotations.

The goal of this paper is to apply some of the best-known machine learning techniques to a practical problem in the automotive field: the identification and classification of the user's intentions in performing specific driving maneuvers. Data have been collected by a static driving simulator. These models are then analyzed and compared, in order to select the best car-following maneuver classifier.

Signal finding (pattern discovery) in biological sequences is a fundamental problem in both computer science and molecular biology. Many approaches have been proposed for extracting interesting patterns (or motifs) from DNA/RNA and protein sequences. Some approaches are based on simple and multiple alignment techniques, some use biological knowledge and others do not.In this paper, we propose a de novo framework that performs motifs identification and exploits a constrained co-clustering technique allowing one to simultaneously find associations between groups of protein sequences and groups of motifs.We show that the presented approach is able to group together protein sequences belonging to the same families and, at the same time to provide a set of characterizing motifs.

The lack of competition culture is one of the keys factors which undermine the enforcement of the competition law in an emerging economy. The lack of competition culture may only be ‘cured’ through activities of competition advocacy vis-a-vis the different stakeholders: State’s agencies, regulators, business community and consumers’ association. The paper analyzes this broad issue through a case study on the Brazilian experience in the field of competition advocacy. During the last years Brazil sensibly progressed in the enforcement of the competition law, and one of the reasons of this development was the functions of competition advocacy carried out by the Brazilian competition authority. The paper argues that the Brazilian success was mainly due to the institutional framework provided by the Brazilian competition act, which splits the activities of enforcement and of competition advocacy between two different agencies. While a fully independent authority enforces the competition law, an advisory body part of the Ministry of the Economy monitors the anti-competitive regulations adopted by other State agencies. The conclusions provide a number of lessons applicable to other emerging economies which have recently started to enforce a competition law regime.

In recent years, researchers have begun to study inductive databases, a new generation of databases for leveraging decision support applications. In this context, the user interacts with the DBMS using advanced, constraint-based languages for data mining where constraints have been specifically introduced to increase the relevance of the results and, at the same time, to reduce its volume. In this paper we study the problem of mining frequent itemsets using an inductive database. We propose a technique for query answering which consists in rewriting the query in terms of union and intersection of the result sets of other queries, previously executed and materialized. Unfortunately, the exploitation of past queries is not always applicable. We then present sufficient conditions for the optimization to apply and show that these conditions are strictly connected with the presence of functional dependencies between the attributes involved in the queries. We show some experiments on an initial prototype of an optimizer which demonstrates that this approach to query answering is viable and in many practical cases it drastically reduces the query execution time.

In recent years, the KDD process has been advocated to be an iterative and interactive process. It is seldom the case that a user is able to answer immediately with a single query all his questions on data. On the contrary, the workflow of the typical user consists in several steps, in which he/she iteratively refines the extracted knowledge by inspecting previous results and posing new queries. Given this view of the KDD process, it becomes crucial to have KDD systems that are able to exploit past results thus minimizing computational effort. This is expecially true in environments in which the system knowledge base is the result of many discoveries on data made separately by the collaborative effort of different users. In this paper, we consider the problem of mining frequent association rules from database relations. We model a general, constraint-based, mining language for this task and study its properties w.r.t. the problem of re-using past results. In particular, we individuate two class of query constraints, namely “item dependent” and “context dependent” ones, and show that the latter are more difficult than the former ones. Then, we propose two newly developed algorithms which allow the exploitation of past results in the two cases. Finally, we show that the approach is both effective and viable by experimenting on some datasets.

A number of techniques have been developed in order to address issues such as genome, trascriptome and proteome analysis. However, a time and cost effective technique for interactome analysis is still lacking. Lots of methods for the predicion of protein - protein interacions have been developed: some of them are based on high quality alignment of sequences, others are based on the tridimensional features of proteins, but they all bear strong limitations that make impossible their large scale application. Recently, an SVM-based machine learning approach has been used to address this topic. Although the method was able to correctly classify 80% of the test samples, it was not applied to the prediction of yet unknown interactions. In this work, we address this topic and show that an optimized, SVM-based machine learning approach trained with combinations of shuffled sequences as examples of lack of interaction is unable to make large scale predictions of interaction. © Springer-Verlag Berlin Heidelberg 2003.

The multilineage hematological effects of the chemoprotective agent amifostine suggested administration of the drug to patients with myelodysplastic syndromes. Partial but encouraging responses were seen. The aim of this research was to further evaluate amifostine activity on the hemopoietic progenitor cells of myelodysplastic patients. We determined amifostine effects on in vitro growth of pre-treatment peripheral blood and bone marrow hemopoietic progenitor cells from myelodysplastic patients, and colony forming unit granulocyte-macrophage and burst forming unit erythroid levels in the peripheral blood of 7 myelodysplastic patients during and following a 3-week treatment with amifostine (200 mg/m2 i.v., three times a week). No variations were observed for hemoglobin, platelets, white blood cells and reticulocyte values. The peripheral blood levels of both colony forming unit granulocyte-macrophage and burst forming unit erythroid, however, showed a significant and sustained increase. In vitro a significant effect exerted by the drug was observed in single cases, but the in vitro effect did not predict the in vivo increase of peripheral blood levels of hemopoietic progenitor cells. These results appear to provide evidence that amifostine has a significant impact on early events of hemopoiesis in myelodysplastic patients.

Recently inductive databases (IDBs) have been proposed to afford the problem of knowledge discovery from huge databases. With an IDB the user/analyst performs a set of very different operations on data using a special-purpose language, powerful enough to perform all the required manipulations, such as data preprocessing, pattern discovery and pattern post-processing. In this paper we present a comparison between query languages (MSQL, DMQL and MINE RULE) that have been proposed for association rules extraction in the last years and discuss their common features and differences. We present them using a set of examples, taken from the real practice of data mining. This allows us to define the language design guidelines, with particular attention to the open issues on IDBs.

This paper focuses on a major step of machine learning, namely checking whether an example matches a candidate hypothesis. In relational learning, matching can be viewed as a Constraint Satisfaction Problem (CSP). The complexity of the task is analyzed in the Phase Transition framework, investigating the impact on the effectiveness of two relational learners: FOIL and G-NET. The critical factors of complexity, and their critical values, are experimentally investigated on artificial problems. This leads to distinguish several kinds of learning domains, depending on whether the target concept lies in the "mushy" region or not. Interestingly, experiments done with FOIL and G-NET show that both learners tend to induce hypotheses generating matching problems located inside the phase transition region, even if the constructed target concept lies far outside. Moreover, target concepts constructed too close to the phase transition are hard and both learners fail. The paper offers an explanation for this fact, and proposes a classification of learning domains and their hardness.

This paper presents the experimental evaluation of a neural network architecture that can manage structured data and refine knowledge bases expressed in a first order logic language.

This paper presents an experimental analysis of a method recently proposed for refining knowledge bases expressed in a first order logic language. The method consists in transforming a classification theory into a neural network, called First Order logic Neural Network (FONN), by replacing predicate semantic functions and logical connectives with continuous-valued derivable functions. In this way it is possible to tune numerical constants in the original theory by performing the error gradient descent. The classification theory to be refined can be manually handcrafted or automatically acquired by a symbolic relational learning system able to deal with numerical features. The emphasis of this paper is on the experimental analysis of the method and an extensive experimentation is provided considering different choices for encoding the logical connectives, and different variants of the learning strategy. The experimentation is made on a challenging artificial case study and shows that FONNs converge quite fastly and generalize better than propositional learners do on an equivalent task definition.

The symbolic approach to machine learning has developed algorithms for learning first-order logic concept definitions. Nevertheless, most of them are limited because of their inability to cope with numeric features, typical of real-world data. A method to overcome this problem is proposed. In particular, an extended version of the system ML-SMART is described, which is capable of automatically adjusting the values of fuzzy sets used to define the semantics of the predicates in the concept description language. The learning strategy works in two separate phases. In the first phase, the structure of the concept definition is learned by choosing tentative values for the fuzzy sets. In the second phase, the values are refined using a simple genetic algorithm by trying to get closer to an optimum assignment. The system is evaluated on a complex artificial domain that shows the good potentialities of this approach. >

Inducing concept descriptions in first order logic is inherently a complex task; then, heuristics are needed to keep the problem to manageable size. In this paper we explore the effect of alternative search strategies, including the use of information gain and of a-priori knowledge, on the quality of the acquired relations, intended as the ability to reconstruct the rule used to generate the examples. To this aim, an artificial domain has been created, in which the experimental conditions can be kept under control, the “solulion” of the learning problem is known and a perfect theory is available. Another investigated aspect is the impact of more complex description languages, such as, for instance, including numerical quantifiers. The resultS show that the information gain criterion is too greedy to be useful when the concepts have a complex internal structure; however, this drawback is more or less shared with any purely statistical evaluation criterion. The addition of parts of the available domain theory increases the obtained performance level. Similar results have been previously obtained on a number of real applications and of test-cases taken from standard machine learning data bases.

This article presents the system WHY, which learns and updates a diagnostic knowledge base using domain knowledge and a set of examples. The a priori knowledge consists of a causal model of the domain that states the relationships among basic phenomena, and a body of phenomenological theory that describes the links between abstract concepts and their possible manifestations in the world. The phenomenological knowledge is used deductively, the causal model is used abductively, and the examples are used inductively. The problems of imperfection and intractability of the theory are handled by allowing the system to make assumptions during its reasoning. In this way, robust knowledge can be learned with limited complexity and a small number of examples. The system works in a first-order logic environment and has been applied in a real domain.

In this paper we briefly survey the problems arising in learning concept descriptions from examples in domains affected by uncertainty and vagueness. A programming environment, called SMART-SHELL, is also presented: it addresses these problems, exploiting fuzzy logic. This is achieved by supplying the learning system with the capability of handling a fuzzy relational database, containing the extensional representation of the acquired logic formulas.