Search

International audience; Humanity is facing existential, societal challenges related to food security, ecosystem conservation, antimicrobial resistance, etc, and Artificial Intelligence (AI) is already playing an important role in tackling these new challenges. Most current AI approaches are limited when it comes to ‘knowledge transfer’ with humans, i.e. it is difficult to incorporate existing human knowledge and also the output knowledge is not human comprehensible. In this chapter we demonstrate how a combination of comprehensible machine learning, text-mining and domain knowledge could enhance human-machine collaboration for the purpose of automated scientific discovery where humans and computers jointly develop and evaluate scientific theories. As a case study, we describe a combination of logic-based machine learning (which included human-encoded ecological background knowledge) and text-mining from scientific publications (to verify machine-learned hypotheses) for the purpose of automated discovery of ecological interaction networks (food-webs) to detect change in agricultural ecosystems using the Farm Scale Evaluations (FSEs) of genetically modified herbicide-tolerant (GMHT) crops dataset. The results included novel food-web hypotheses, some confirmed by subsequent experimental studies (e.g. DNA analysis) and published in scientific journals. These machine-leaned food-webs were also used as the basis of a recent study revealing resilience of agro-ecosystems to changes in farming management using GMHT crops.

International audience; There is growing interest in the potential for combining eDNA and artificial intelligence (machine learning) to detect and evaluate in real time changes in ecosystems at the global scale, in a more sensitive and cost-effective way than current biomonitoring methods. Machine learning might make better use of the eDNA census data that can currently be collected to evaluate the network of ecological interactions that are at the base of the services that ecosystems supply and that we wish to protect. To date, eDNA and machine learning developments have effectively progressed in parallel and in isolation in various spheres of ecosystem monitoring (disease, invasion, conservation, etc. in aerial, terrestrial, and aquatic systems).The goal of this Research Topic is to explore the range of ongoing activities to build the next generation of biomonitoring tools and in doing so to make researchers in the different spheres aware of the breadth of work being undertaken, and to set a unifying research agenda (the key questions) for the development of global biomonitoring using eDNA and machine learning.The scope of this Research Topic will be to explore:1. eDNA approaches currently being used in case study systems from all spheres of monitoring;2. Theoretical underpinnings of machine learning for biomonitoring;3. What type of networks do we need to reconstruct for effective monitoring (co-occurrence, trophic, etc);4. Examples of learning large scale, replicated networks from eDNA in the different spheres;5. Statistical and analytical approaches to analysing large-scale, highly replicated networks;6. Technological developments necessary to build a next-generation biomonitoring framework at the global scale;7. A research agenda paper that develops “10 key questions for eDNA and machine learning in biomonitoring”.Details for Authors: The Research Topic “A next-generation of global biomonitoring to detect ecosystem change” will publish conceptual, data, case study, technological and synthetic papers on eDNA and machine learning approaches for developing a unified next-generation biomonitoring framework. Paper length conforms to the guidelines of the journal Frontiers in Ecology and Evolution.

Statistical machine learning is widely used in image classification. However, most techniques (1) require many images to achieve high accuracy and (2) do not provide support for reasoning below the level of classification, and so are unable to support secondary reasoning, such as the existence and position of light sources and other objects outside the image. This paper describes an Inductive Logic Programming approach called Logical Vision which overcomes some of these limitations. LV uses Meta-Interpretive Learning (MIL) combined with low-level extraction of high-contrast points sampled from the image to learn recursive logic programs describing the image. In published work LV was demonstrated capable of high-accuracy prediction of classes such as regular polygon from small numbers of images where Support Vector Machines and Convolutional Neural Networks gave near random predictions in some cases. LV has so far only been applied to noise-free, artificially generated images. This paper extends LV by (a) addressing classification noise using a new noise-telerant version of the MIL system Metagol, (b) addressing attribute noise using primitive-level statistical estimators to identify sub-objects in real images, (c) using a wider class of background models representing classical 2D shapes such as circles and ellipses, (d) providing richer learnable background knowledge in the form of a simple but generic recursive theory of light reflection. In our experiments we consider noisy images in both natural science settings and in a RoboCup competition setting. The natural science settings involve identification of the position of the light source in telescopic and microscopic images, while the RoboCup setting involves identification of the position of the ball. Our results indicate that with real images the new noise-robust version of LV using a single example (i.e. one-shot LV) converges to an accuracy at least comparable to a thirty-shot statistical machine learner on both prediction of hidden light sources in the scientific settings and in the RoboCup setting. Moreover, we demonstrate that a general background recursive theory of light can itself be invented using LV and used to identify ambiguities in the convexity/concavity of objects such as craters in the scientific setting and partial obscuration of the ball in the RoboCup setting.

Sustainable management of ecosystems and growth in agricultural productivity is at the heart of the United Nations' Sustainable Development Goals for 2030. New management regimes could revolutionize agricultural production, but require an evaluation of the risks and opportunities. Replacing existing conventional weed management with genetically modified, herbicide-tolerant (GMHT) crops, for example, might reduce herbicide applications and increase crop yields, but remains controversial owing to concerns about potential impacts on biodiversity. Until now, such new regimes have been assessed at the species or assemblage level, whereas higher-level ecological network effects remain largely unconsidered. Here, we conduct a large-scale network analysis of invertebrate communities across 502 UK farm sites to GMHT management in different crop types. We find that network-level properties were overwhelmingly shaped by crop type, whereas network structure and robustness were apparently unaltered by GMHT management. This suggests that taxon-specific effects reported previously did not escalate into higher-level systemic structural change in the wider agricultural ecosystem. Our study highlights current limitations of autecological assessments of effect in agriculture in which species interactions and potential compensatory effects are overlooked. We advocate adopting the more holistic system-level evaluations that we explore here, which complement existing assessments for meeting our future agricultural needs.

During the 1980s Michie defined Machine Learning in terms of two orthogonal axes of performance: predictive accuracy and comprehensibility of generated hypotheses. Since predictive accuracy was readily measurable and comprehensibility not so, later definitions in the 1990s, such as Mitchell’s, tended to use a one-dimensional approach to Machine Learning based solely on predictive accuracy, ultimately favouring statistical over symbolic Machine Learning approaches. In this paper we provide a definition of comprehensibility of hypotheses which can be estimated using human participant trials. We present two sets of experiments testing human comprehensibility of logic programs. In the first experiment we test human comprehensibility with and without predicate invention. Results indicate comprehensibility is affected not only by the complexity of the presented program but also by the existence of anonymous predicate symbols. In the second experiment we directly test whether any state-of-the-art ILP systems are ultra-strong learners in Michie’s sense, and select the Metagol system for use in humans trials. Results show participants were not able to learn the relational concept on their own from a set of examples but they were able to apply the relational definition provided by the ILP system correctly. This implies the existence of a class of relational concepts which are hard to acquire for humans, though easy to understand given an abstract explanation. We believe improved understanding of this class could have potential relevance to contexts involving human learning, teaching and verbal interaction.

Statistical machine learning is widely used in image classification. However, most techniques (1) require many images to achieve high accuracy and (2) do not provide support for reasoning below the level of classification, and so are unable to support secondary reasoning, such as the existence and position of light sources and other objects outside the image. In recent work an Inductive Logic Programming approach called Logical Vision (LV) was shown to overcome some of these limitations. LV uses Meta-Interpretive Learning combined with low-level extraction of high-contrast points sampled from the image to learn recursive logic programs describing the image. This paper extends LV by using (a) richer background knowledge enabling secondary reasoning from raw images, such as light reflection that can itself be learned and used for resolving visual ambiguities, which cannot be easily modelled using statistical approaches, (b) a wider class of background models representing classical 2D shapes such as circles and ellipses, (c) primitive-level statistical estimators to handle noise in real images. Our results indicate that the new noise-robust version of LV is able to handle secondary reasoning task in real images with few data, which is very similar to scientific discovery process of humans. Specifically, it uses a single example (i.e. one-shot LV) converges to an accuracy at least comparable to thirty-shot statistical machine learner on the prediction of hidden light sources. Moreover, we demonstrate that the learned theory can be used to identify ambiguities in the convexity/concavity of objects such as craters.

During the 1980s Michie defined Machine Learning in terms of two orthogonal axes of performance: predictive accuracy and comprehensibility of generated hypotheses. Since predictive accuracy was readily measurable and comprehensibility not so, later definitions in the 1990s, such as that of Mitchell, tended to use a one-dimensional approach to Machine Learning based solely on predictive accuracy, ultimately favouring statistical over symbolic Machine Learning approaches. In this paper we provide a definition of comprehensibility of hypotheses which can be estimated using human participant trials. We present the results of experiments testing human comprehensibility of logic programs learned with and without predicate invention. Results indicate that comprehensibility is affected not only by the complexity of the presented program but also by the existence of anonymous predicate symbols.

Despite early interest Predicate Invention has lately been under-explored within ILP. We develop a framework in which predicate invention and recursive generalisations are implemented using abduction with respect to a meta-interpreter. The approach is based on a previously unexplored case of Inverse Entailment for Grammatical Inference of Regular languages. Every abduced grammar H is represented by a conjunction of existentially quantified atomic formulae. Thus ¬H is a universally quantified clause representing a denial. The hypothesis space of solutions for ¬H can be ordered by ?-subsumption. We show that the representation can be mapped to a fragment of Higher-Order Datalog in which atomic formulae in H are projections of first-order definite clause grammar rules and the existentially quantified variables are projections of first-order predicate symbols. This allows predicate invention to be effected by the introduction of first-order variables. Previous work by Inoue and Furukawa used abduction and meta-level reasoning to invent predicates representing propositions. By contrast, the present paper uses abduction with a meta-interpretive framework to invent relations. We describe the implementations of Meta-interpretive Learning (MIL) using two different declarative representations: Prolog and Answer Set Programming (ASP). We compare these implementations against a state-of-the-art ILP system MC-TopLog using the dataset of learning Regular and Context-Free grammars as well learning a simplified natural language grammar and a grammatical description of a staircase. Experiments indicate that on randomly chosen grammars, the two implementations have significantly higher accuracies than MC-TopLog. In terms of running time, Metagol is overall fastest in these tasks. Experiments indicate that the Prolog implementation is competitive with the ASP one due to its ability to encode a strong procedural bias. We demonstrate that MIL can be applied to learning natural grammars. In this case experiments indicate that increasing the available background knowledge, reduces the running time. Additionally ASPM (ASP using a meta-interpreter) is shown to have a speed advantage over Metagol when background knowledge is sparse. We also demonstrate that by combining MetagolR (Metagol with a Regular grammar meta-interpreter) and MetagolCF (Context-Free meta-interpreter) we can formulate a system, MetagolRCF, which can change representation by firstly assuming the target to be Regular, and then failing this, switch to assuming it to be Context-Free. MetagolRCF runs up to 100 times faster than MetagolCF on grammars chosen randomly from Regular and non-Regular Context-Free grammars.

We foresee a new global-scale, ecological approach to biomonitoring emerging within the next decade that can detect ecosystem change accurately, cheaply, and generically. Next-generation sequencing of DNA sampled from the Earth's environments would provide data for the relative abundance of operational taxonomic units or ecological functions. Machine-learning methods would then be used to reconstruct the ecological networks of interactions implicit in the raw NGS data. Ultimately, we envision the development of autonomous samplers that would sample nucleic acids and upload NGS sequence data to the cloud for network reconstruction. Large numbers of these samplers, in a global array, would allow sensitive automated biomonitoring of the Earth's major ecosystems at high spatial and temporal resolution, revolutionising our understanding of ecosystem change.

Data transformation involves the manual construction of large numbers of special-purpose programs. Although typically small, such programs can be complex, involving problem decomposition, recursion, and recognition of context. Building such programs is common in commercial and academic data analytic projects and can be labour intensive and expensive, making it a suitable candidate for machine learning. In this paper, we use the meta-interpretive learning framework (MIL) to learn recursive data transformation programs from small numbers of examples. MIL is well suited to this task because it supports problem decomposition through predicate invention, learning recursive programs, learning from few examples, and learning from only positive examples. We apply Metagol, a MIL implementation, to both semi-structured and unstructured data. We conduct experiments on three real-world datasets: medical patient records, XML mondial records, and natural language taken from ecological papers. The experimental results suggest that high levels of predictive accuracy can be achieved in these tasks from small numbers of training examples, especially when learning with recursion.

In a previous paper we described a machine learning approach which was used to automatically generate food-webs from national-scale agricultural data. The learned food-webs in the previous study consist of hundreds of ground facts representing trophic links between individual species. These species food-webs can be used to explain the structure and dynamics of particular eco-systems, however, they cannot be directly used as general predictive models. In this paper we describe the first steps towards this generalisation and present initial results on i learning general functional food-webs i.e. trophic links between functional groups of species and ii meta-interpretive learning MIL of general predictive rules e.g. about the effect of agricultural management. Experimental results suggest that functional food-webs have at least the same levels of predictive accuracies as species food-webs despite being much more compact. In this paper we also present initial experiments where predicate invention and recursive rule learning in MIL are used to learn food-webs as well as predictive rules directly from data.

In this paper we use a logic-based representation and a combination of Abduction and Induction to model inhibition in metabolic networks. In general, the integration of abduction and induction is required when the following two conditions hold. Firstly, the given background knowledge is incomplete. Secondly, the problem must require the learning of general rules in the circumstance in which the hypothesis language is disjoint from the observation language. Both these conditions hold in the application considered in this paper. Inhibition is very important from the therapeutic point of view since many substances designed to be used as drugs can have an inhibitory effect on other enzymes. Any system able to predict the inhibitory effect of substances on the metabolic network would therefore be very useful in assessing the potential harmful side-effects of drugs. In modelling the phenomenon of inhibition in metabolic networks, background knowledge is used which describes the network topology and functional classes of inhibitors and enzymes. This background knowledge, which represents the present state of understanding, is incomplete. In order to overcome this incompleteness hypotheses are considered which consist of a mixture of specific inhibitions of enzymes (ground facts) together with general (non-ground) rules which predict classes of enzymes likely to be inhibited by the toxin. The foreground examples are derived from in vivo experiments involving NMR analysis of time-varying metabolite concentrations in rat urine following injections of toxins. The model's performance is evaluated on training and test sets randomly generated from a real metabolic network. It is shown that even in the case where the hypotheses are restricted to be ground, the predictive accuracy increases with the number of training examples and in all cases exceeds the default (majority class). Experimental results also suggest that when sufficient training data is provided, non-ground hypotheses show a better predictive accuracy than ground hypotheses. The model is also evaluated in terms of the biological insight that it provides. 64 1-3 209 230 Cited By :53

International audience; Networks of trophic links (food webs) are used to describe and understand mechanistic routes for translocation of energy (biomass) between species. However, a relatively low proportion of ecosystems have been studied using food web approaches due to difficulties in making observations on large numbers of species. In this paper we demonstrate that Machine Learning of food webs, using a logic-based approach called A/ILP, can generate plausible and testable food webs from field sample data. Our example data come from a national-scale Vortis suction sampling of invertebrates from arable fields in Great Britain. We found that 45 invertebrate species or taxa, representing approximately 25% of the sample and about 74% of the invertebrate individuals included in the learning, were hypothesized to be linked. As might be expected, detritivore Collembola were consistently the most important prey. Generalist and omnivorous carabid beetles were hypothesized to be the dominant predators of the system. We were, however, surprised by the importance of carabid larvae suggested by the machine learning as predators of a wide variety of prey. High probability links were hypothesized for widespread, potentially destabilizing, intra-guild predation; predictions that could be experimentally tested. Many of the high probability links in the model have already been observed or suggested for this system, supporting our contention that A/ILP learning can produce plausible food webs from sample data, independent of our preconceptions about "who eats whom." Well-characterised links in the literature correspond with links ascribed with high probability through A/ILP. We believe that this very general Machine Learning approach has great power and could be used to extend and test our current theories of agricultural ecosystem dynamics and function. In particular, we believe it could be used to support the development of a wider theory of ecosystem responses to environmental change.

Worldwide demand for food will increase dramatically in the future as global human population grows. Increasing efficiency of crop production is unlikely to be sufficient to meet the demand, presenting a long-term threat to humanity’s ‘well-being’. Knowledge of the system-level behaviour of agroecosystems, however, remains surprisingly limited, reflecting the agricultural focus on particular species. This is starting to change towards an ecosystem and network-based approach, following the recent revolution in thinking about resource use and sustainability in our other global food production industry: commercial fisheries. Agroecosystems appear to retain plasticity of ecological processes that might be manipulated for productivity and sustainability. Network structure and dynamics have substantial impacts on ecosystem performance, but evidence from agroecosystems lags behind network theory. Here, we provide an introduction to network theory and application in agroecosystems, identify network metrics for management and environmental change, and, finally, we highlight gaps in our current knowledge and key research themes. These themes include: is the structure of agroecological networks affected by sampling; how do ecosystem services ‘emerge’ empirically from ecological organization, function and network properties; how do spatial and temporal scale and resolution influence system performance; and, can network agroecology be used to design systems that maximize ecosystem services?

Chapitre 4; International audience; Network ecology holds great promise as an approach to modelling and predicting the effects of agricultural management on ecosystem service provision, as it bridges the gap between community and ecosystem ecology. Unfortunately, trophic interactions between most species in agricultural farmland are not well characterised empirically, and only partial food webs are available for a few systems. Large agricultural datasets of the nodes (i.e., species) in the webs are now available, and if these can be enriched with information on the links between them then the current shortage of network data can potentially be overcome. We demonstrate that a logic-based machine learning method can be used to automatically assign interactions between nodes, thereby generating plausible and testable food webs from ecological census data. Many of the learned trophic links were corroborated by the literature: in particular, links ascribed with high probability by machine learning corresponded with those having multiple references in the literature. In some cases, previously unobserved but high probability links were suggested and subsequently confirmed by other research groups. We evaluate these food webs using a new cross-validation method and present new results on automatic corroboration of a large, complex food web. The simulated frequencies of trophic links were also correlated with the total number of literature 'hits' for these links from the automatic corroboration. Finally, we also show that a network constructed by learning trophic links between functional groups is at least as accurate as the species-based trophic network.

International audience; Machine Learning has been used to automatically generate a probabilistic food-web from Farm Scale Evaluation (FSE) data. The initial food web proposed by machine learning has been examined by domain experts and comparison with the literature shows that many of the links are corroborated. The FSE data were collected using two different sampling techniques, namely Vortis and pitfall. The corroboration of the initial Vortis food web, generated by machine learning, was performed manually by the domain experts. However, manual corroboration of hypothetical trophic links is difficult and requires significant amounts of time. In this paper we review the method and the main results on machine learning of trophic links. We study common trophic links from Vortis and pitfall data. We also describe a new method and present initial results on automatic corroboration of trophic links using text mining.

International audience; In this paper we demonstrate that machine learning (using Abductive ILP) can generate plausible and testable food webs from ecological data. In this approach, unlike previous applications of Abductive ILP, the abductive predicate ‘eats’ is entirely undefined before the start of the learning. We also explore a new approach, called Hypothesis Frequency Estimation (HFE), for estimating probabilities for hypothetical ‘eats’ facts based on their frequency of occurrence when randomly sampling the hypothesis space. The results of cross-validation tests suggest that the trophic networks with probabilities have higher predictive accuracies compared to the networks without probabilities. The proposed trophic networks have been examined by domain experts and comparison with the literature shows that many of the links are corroborated by the literature. In particular, links ascribed with high frequency are shown to correspond well with those having multiple references in the literature. In some cases novel high frequency links are suggested, which could be tested.

Within ILP much effort has been put into designing methods that are complete for hypothesis finding. However, it is not clear whether completeness is important in real-world applications. This paper uses a simplified version of grammar learning to show how a complete method can improve on the learning results of an incomplete method. Seeing the necessity of having a complete method for real-world applications, we introduce a method called ⊤-directed theory co-derivation, which is shown to be correct (ie. sound and complete). The proposed method has been implemented in the ILP system MC-TopLog and tested on grammar learning and the learning of game strategies. Compared to Progol5, an efficient but incomplete ILP system, MC-TopLog has higher predictive accuracies, especially when the background knowledge is severely incomplete.

This paper describes the use of a mixture of abduction and induction for the temporal modelling of the effects of toxins in metabolic networks. Background knowledge is used which describes network topology and functional classes of enzymes. This background knowledge, which represents the present state of understanding, is incomplete. In order to overcome this incompleteness hypotheses are considered which consist of a mixture of specific inhibitions of enzymes (ground facts) together with general (non-ground) rules which predict classes of enzymes likely to be inhibited by the toxin. The foreground examples were derived from in vivo experiments involving NMR analysis of time-varying metabolite concentrations in rat urine following injections of toxin. Hypotheses about inhibition are built using the Inductive Logic Programming system Progo15.0 and predictive accuracy is assessed for both the ground and the non-ground cases. © 2005 IEEE. 2005 233 238 Sponsors: California State University Bakersfield and Association for Machine Learning and Applications Conference code: 69275 Cited By :3

In a previous paper we introduced a framework for combining Genetic Algorithms with ILP which included a novel representation for clauses and relevant operators. In this paper we complete the proposed framework by introducing a fast evaluation mechanism. In this evaluation mechanism individuals can be evaluated at genotype level (i.e. bit-strings) without mapping them into corresponding clauses. This is intended to replace the complex task of evaluating clauses (which usually needs repeated theorem proving) with simple bitwise operations. In this paper we also provide an experimental evaluation of the proposed framework. The results suggest that this framework could lead to significantly increased efficiency in problems involving complex target theories.

Bioinformatics is characterised by a growing diversity of large-scale databases containing information on genetics, proteins, metabolism and disease. It is widely agreed that there is an increasingly urgent need for technologies which can integrate these disparate knowledge sources. In this paper we propose that not only is machine learning a good candidate technology for such data integration, but Inductive Logic Programming, in particular, has strengths for handling the relational aspects of this task. Relations can be used to capture, in a single representation, not only biochemical reaction information but also protein and ligand structure as well as metabolic network information. Resources such as the Gene Ontology (GO) and the Enzyme Commission (EC) system both provide isa-hierarchies of enzyme functions. On the face of it GO and EC should be invaluable resources for supporting automation within Functional Genomics, which aims at predicting the function of unassigned enzymes from the genome projects. However, neither GO nor EC can be directly used for this purpose since the classes have only a natural language description. In this paper we make an initial attempt at machine learning EC classes for the purpose of enzyme function prediction in terms of biochemical reaction descriptions found in the LIGAND database. To our knowledge this is the first attempt to do so. In our experiments we learn descriptions for a small set of EC classes including Oxireductase and Phosphotransferase. Predictive accuracy are provided for all learned classes. In further work we hope to complete the learning of enzyme classes and integrate the learned models with metabolic network descriptions to support “gap-filling” in the present understanding of metabolism.

In this paper, we study how a logical form of scientific modelling that integrates together abduction and induction can be used to understand the functional class of unknown enzymes or inhibitors. We show how we can model, within Abductive Logic Programming (ALP), inhibition in metabolic pathways and use abduction to generate facts about inhibition of enzymes by a particular toxin (e.g. Hydrazine) given the underlying metabolic pathway and observations about the concentration of metabolites. These ground facts, together with biochemical background information, can then be generalised by ILP to generate rules about the inhibition by Hydrazine thus enriching further our model. In particular, using Progol 5.0 where the processes of abduction and inductive generalization are integrated enables us to learn such general rules. Experimental results on modelling in this way the effect of Hydrazine in a real metabolic pathway are presented.

Logic programs with annotated disjunctions, or LPADs, are an elegant knowledge representation formalism that can be used to combine first order logical and probabilistic inference. While LPADs can be written manually, one can also consider the question of how to learn them from data. Methods for learning restricted classes of LPADs have been proposed before, but the problem of learning any kind of LPADs was still open. In this paper, we describe a reduction of non-recursive LPADs with a finite Herbrand universe to Bayesian networks. This reduction makes it possible to learn such LPADs using standard learning techniques for Bayesian networks. Thus the class of learnable LPADs is extended. © Springer-Verlag Berlin Heidelberg 2007. ispartof: pages:94-108 ispartof: Inductive Logic Programming, ILP 2006, Revised Selected Papers vol:4455 pages:94-108 ispartof: Inductive Logic Programming location:Santiago de Compostela, Spain date:24 Aug - 27 Aug 2006 status: published

The discovery of relationships between concepts is a crucial point in ontology learning (OL). In most cases, OL is achieved from a collection of domain-specific texts, describing the concepts of the domain and their relationships. A natural way to represent the description associated to a particular text is to use a structured term (or tree). We present a method for learning transformation rules, rewriting natural language texts into trees, where the input examples are couples (text, tree). The learning process produces an ordered set of rules such that, applying these rules to a textgives the corresponding tree.

Automatically extracting spatial information is a challenging novel task with many applications. We formalize it as an information extraction step required for a mapping from natural language to formal spatial representation. Sentences may give rise to multiple spatial realtions between words representing landmarks, trajectors and spatial indicators. Our contribution is to formulate the extraction task as a relational learning problem, for which we employ the recently introduced kLog framework. We discuss representational and modeling aspects, kLog’s flexibility in our task and we present current experimental results. ispartof: pages:204-220 ispartof: Lecture notes in artificial intelligence vol:7207 pages:204-220 ispartof: Inductive Logic Programming location:Windsor status: published

Hedge cue detection is a Natural Language Processing (NLP) task that consists of determining whether sentences contain hedges. These linguistic devices indicate that authors do not or cannot back up their opinions or statements with facts. This binary classification problem, i.e. distinguishing factual versus uncertain sentences, only recently received attention in the NLP community. We use kLog, a new logical and relational language for kernel-based learning, to tackle this problem. We present results on the CoNLL 2010 benchmark dataset that consists of a set of paragraphs from Wikipedia, one of the domains in which uncertainty detection has become important. Our approach shows competitive results compared to state-of-the-art systems. ispartof: pages:347-357 ispartof: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) vol:7207 LNAI pages:347-357 ispartof: Meeting of Computational Linguistics in The Netherlands (CLIN) location:Tilburg, The Netherlands date:20 Jan - 20 Jan 2012 status: published

Model trees are a special case of regression trees in which linear regression models are constructed in the leaves. Little attention has been paid to model trees in relational learning, mainly because the task of learning linear regression equations in this context involves dealing with non-determinacy of predictive attributes. Whereas existing approaches handle this non-determinacy issue either by selecting a single value or by aggregating over all values, in this paper we present a model tree learning system that combines both. © Springer-Verlag Berlin Heidelberg 2007. ispartof: pages:424-438 ispartof: Lecture notes in computer science vol:4455 pages:424-438 ispartof: Inductive Logic Programming, ILP 2006 location:Santiago de Compostela, Spain date:24 Aug - 27 Aug 2006 status: published

Recently, there has been an increasing interest in directed probabilistic logical models and a variety of languages for describing such models has been proposed. Although many authors provide high-level arguments to show that in principle models in their language can be learned from data, most of the proposed learning algorithms have not yet been studied in detail. We introduce an algorithm, generalized ordering-search, to learn both structure and conditional probability distributions (CPDs) of directed probabilistic logical models. The algorithm upgrades the ordering-search algorithm for Bayesian networks. We use relational probability trees as a representation for the CPDs. We present experiments on blocks world domains, a gene domain and the Cora dataset. ispartof: pages:40-42 ispartof: Lecture Notes in Computer Science vol:4455 pages:40-42 ispartof: Inductive Logic Programming location:Santiago de Compostella, Spain date:24 Aug - 27 Aug 2006 status: published