Search

We evaluated probabilistic lexicalized tree-insertion grammars (PLTIGs) on a classification task relevant for automatic speech recognition. The baseline is a family of n-gram models tuned with Witten-Bell smoothing. The language models are trained on unannotated corpora, consisting of 10,000 to 50,000 sentences collected from the English section of Wikipedia. For the evaluation, an additional 150 random sentences were selected from the same source, and for each of these, approximately 3,200 variations were generated. Each variant sentence was obtained by replacing an arbitrary word by a similar word, chosen to be at most 2 character edits from the original. The evaluation task consisted of identifying the original sentence among the automatically constructed (and typically inferior) alternatives. In the experiments, the n-gram models outperformed the PLTIG model on the smaller data set, but as the size of data grew, the PLTIG model gave comparable results. While PLTIGs are more demanding to train, they have the advantage that they assign a parse structure to their input sentences. This is valuable for continued algorithmic processing, for example, for summarization or sentiment analysis.

The efficacy of syntactic features for topic-independent authorship attribution is evaluated, taking a feature set of frequencies of words and punctuation marks as baseline. The features are 'deep' in the sense that they are derived by parsing the subject texts, in contrast to 'shallow' syntactic features for which a part-of-speech analysis is enough. The experiments are made on two corpora of online texts and one corpus of novels written around the year 1900. The classification tasks include classical closed-world authorship attribution, identification of separate texts among the works of one author, and cross-topic authorship attribution. In the first tasks, the feature sets were fairly evenly matched, but for the last task, the syntax-based feature set outperformed the baseline feature set. These results suggest that, compared to lexical features, syntactic features are more robust to changes in topic.

We introduce and investigate stack transducers, which are one-way stack automata with an output tape. A one-way stack automaton is a classical pushdown automaton with the additional ability to move the stack head inside the stack without altering the contents. For stack transducers, we distinguish between a digging and a non-digging mode. In digging mode, the stack transducer can write on the output tape when its stack head is inside the stack, whereas in non-digging mode, the stack transducer is only allowed to emit symbols when its stack head is at the top of the stack. These stack transducers have a motivation from natural-language interface applications, as they capture long-distance dependencies in syntactic, semantic, and discourse structures. We study the computational capacity for deterministic digging and non-digging stack transducers, as well as for their non-erasing and checking versions. We finally show that even for the strongest variant of stack transducers the stack languages are regular.

Industrial meat production has several negative environmental effects. Governments’ agricultural policies aim for cost efficiency combined with high environmental and animal welfare, which puts farmers in a difficult situation trying to navigate between sometimes contradictory requirements. This paper studies how Swedish pig farmers resolve or cope with conflicting goals in pig farming. We have analysed the regulations governing EU and Swedish pig farming. We have also interviewed five Swedish pig farmers about their views of the different goals of pig farming and strategies for resolving conflicts between the goals of low environmental impact, high animal welfare and enough profitability to continue farming. The greatest divide was between the conventional farmers, who emphasized natural resource efficiency, and the organic farmers who stressed animal welfare, multifunctionality and ecosystem service delivery. We suggest four strategies to contribute to resolving some of the conflicting goals: improve communication about different types of pig farming; use public procurement as a driver towards more sustainable pork production; work towards improving the Common Agricultural Policy, perhaps by implementing payments for ecosystem services or multifunctionality; and finally, decrease the total production of pork to lower the emissions per land unit., Funding Agency:Foundation for Baltic and East European Studies

MICO is an open-source platform for cross-media analysis, querying, and recommendation. It is the major outcome of the European research project Media in Context, and has been contributed to by academic and industrial partners from Germany, Austria, Sweden, Italy, and the UK. A central idea is to group sets of related media objects into multimodal content items, and to process and store these as logical units. The platform is designed to be easy to extend and adapt, and this makes it a useful building block for a diverse set of multimedia applications. To promote the platform and demonstrate its potential, we describe our work on a Kaldi-based speech-recognition pipeline.

We propose an algorithm for computing the $N$ best roots of a weighted hypergraph, in which the weight function is given over an idempotent and multiplicatively monotone semiring. We give a set of conditions that ensures that the weight function is well-defined and that solutions exist. Under these conditions, we prove that the proposed algorithm is correct. This generalizes a previous result for weighted tree automata, and in doing so, broadens the practical applications.

We present a taxonomy of algorithms for minimising deterministic bottom-up tree automata (DTAs) over ranked and ordered trees. Automata of this type and its extensions are used in many application areas, including natural language processing (NLP) and code generation. In practice, DTAs can grow very large, but minimisation keeps things manageable. The proposed taxonomy serves as a unifying framework that makes algorithms accessible and comparable, and as a foundation for efficient implementation. Taxonomies of this type are also convenient for correctness and complexity analysis, as results can frequently be propagated through the hierarchy. The taxonomy described herein covers a broad spectrum of algorithms, ranging from novel to well-studied ones, with a focus on computational complexity.

The abundance of digital content requires cost-effective technologies to extract the hidden meaning from media objects. However, current approaches fail to deal with the challenges related to cross-media analysis, metadata publishing, querying and recommendation that are necessary to overcome this challenge. In this paper, we describe the EU project MICO (Media in Context) which aims to provide the necessary technologies based on open-source software (OSS) core components.

In this study in Sweden, six small (<50 ha) and six large farms (>135 ha) participated. The aims of the study were to a) measure differences between small and large farms regarding on-farm landscape heterogeneity, and b) evaluate relations between on-farm landscape heterogeneity and biodiversity, measured as numbers of breeding bird species, bird territories, butterflies, bumblebees, and herbaceous plant species. Sample area of the same size, placed on each farm, was used for the biodiversity assessments and on-farm landscape heterogeneity studies. On-farm landscape heterogeneity was classified with the Shannon-Wiener diversity index. Linear regression was applied to analyze relationships between on-farm landscape heterogeneity and biodiversity indicators. Multivariate regression was used to analyze relations between single bird species and specific on-farm habitats. Small farms had significantly higher on-farm landscape heterogeneity than large farms. Strong positive relations between on-farm landscape heterogeneity and number of breeding birds, butterflies, and herbaceous plant species were found. Total on-farm landscape heterogeneity seems to be more important for bird diversity than do specific landscape elements. The study indicates that, to increase biodiversity, farm size should be taken into consideration., Funding Agencies:Ekhaga FoundationEbba and Sven Schwartz Foundation

One year Master's thesis, Degree of master in Business and Economics, School ofBusiness and Economics at Linnaeus University, Management, 4FE10E SpringSemester 2014.Author: Johanna Björklund and Sarah Gustafsson.Mentor: Magnus Forslund.Title: Lean in the public sector - Potential barriers to a long-term Lean work.Background: Organizational changes are something that is constantly done inorganizations, and in the last decade, there has arisen a plethora of different changemodels. A reorganization that is popular among companies today is Lean. Lean is aconcept that was developed within the Japanese automotive industry, but has evolvedfrom the Japanese shop floor and is now a concept that can be found in, for example,several Swedish municipalities.Purpose: The purpose of this study was to create an understanding of why the conceptLean may have difficulty getting a hold in the in the long term.Methodology: The study was done with a qualitative research approach and through anabductive approach. As we intended to create an understanding, of what potentialbarriers that may affect that a Lean work, in some cases, does not become a long-termcondition, it seemed natural to choose a hermeneutic approach to knowledge. Theempirical data collected via individual unstructured interviews, through a groupinterview, and by non-participant observations. Our collection of theoretical materialconsisted mostly of scientific articles or books written by Academics, because we feltthat the writer that was behind these articles were considered to be reliable. We believethat consultants and other writers stand and fall with Lean's survival and reliabilitytherefore falls.Result, conclusions: That there are some barriers to get Lean-work to become a longtermwork, both we and the theory agree about. The factors we could find at thedepartment Financial assistance and employment was that; there are some resistance to4the concept among the employees, the management has difficulties to "let go ofco, Bakgrund: Organisationsförändringar är något som ständigt görs i organisationer, och under det senaste årtiondet har det uppkommit en uppsjö av olika förändringsmodeller. En organisationsförändring som är populär bland företag idag är Lean. Lean är ett koncept som togs fram inom den japanska bilindustrin, men har utvecklats från det japanska verkstadsgolvet till att idag vara ett koncept som återfinns i bland annat flertalet svenska kommuner. Syfte: Syftet med denna studie var att skapa en förståelse för varför Lean-arbete kan ha svårt att få fäste på lång sikt. Metod: Studien gjordes med en kvalitativ forskningsmetod och genom en abduktiv forskningsansats. Då vi ämnade skapa en förståelse för vilka tänkbara hinder som kan påverka att Lean-arbetet i vissa fall inte blir långsiktigt, så föll sig valet naturligt på att välja en hermeneutisk kunskapssyn. Den empiri som samlats in har gjorts via individuella ostrukturerade intervjuer, via en gruppintervju, samt genom ickedeltagande observationer. Vår insamling av teoretiskt material har mestadels bestått av vetenskapliga artiklar eller böcker skrivna av akademiker, då vi som författare anser att de författare som står bakom dessa artiklar är pålitliga. Vi anser att konsulter och andra författare står och faller med Leans överlevnad och att trovärdigheten därför faller. Resultat, slutsatser: Att det finns vissa hinder för att Lean inte blir långsiktigt är både vi och teorin överens om. De faktorer vi kunde hitta på avdelningen Ekonomiskt bistånd och sysselsättning var bland annat att; Lean ses som motstånd till konceptet, att ledningen har svårt att ” släppa på kontrollen”, bristande medarbetarstyrning och mättnad av konceptet.

Matematiskt begåvade elever är i behov av lika mycket stöd, variation och utmaningar som alla andra elever. Därför är det viktigt att dessa elever inte kommer i skymundan och får ”klara sig själva”. Syftet med denna studie är att synliggöra matematiskt begåvade elevers uppfattning om hur deras matematikundervisning ser ut idag, hur de själva lär sig bäst, samt hur de anser att matematikundervisningen ska utformas för att de ska ha möjlighet att utvecklas efter sin fulla förmåga. Med stöd i teori och tidigare forskning valdes sex matematiskt begåvade elever från år 4 ut för att medverka i denna undersökning. Metoden som valdes var att intervjua dessa sex elever för att få en djupare förståelse för hur de själva uppfattar matematikundervisningen idag, deras egna lärstilar, samt hur de vill arbeta i matematiken för att ha möjlighet utvecklas till sin fulla förmåga. Resultatet från intervjuerna visar att stor del av undervisningen idag genomsyras av ensamarbete och där dessa elever inte får det stöd och den hjälp de behöver. Det visade sig också att dessa elever var väldigt olika i sina lärstilar, precis som alla individer, och därför behöver en varierad undervisning för att alla lärstilar ska inkluderas. Det visade sig att alla eleverna i denna studie söker utmaning, mer arbete i grupp och stöd från lärare för att få möjlighet att utvecklas till sin fulla förmåga., The mathematic gifted pupil needs support, variation and challenge as much as other children. Therefore it is important that these children do not end up in the background and that the teacher does not think that ”they can handle it on their own”. The aims of this study are to explore the mathematic gifted pupils´ beliefs about mathematics teaching, their learning styles and how they want to work to have the opportunity to develop their full mathematical capacity. With support from theory and earlier research we chose six mathematical gifted pupils in the fourth grade to participate in the study. As the research method we chose to interview these pupils in order to get a deeper understanding of how they believe that their own learning in mathematic looks like, how they think they learn the best and how they want to work to have the opportunity to develop their full mathematical capacity. The result and conclusion of the interviews show that according to the pupils, a majority of the learning situations includes the pupils’ individual work and by that these children do not get the help and support that they need. The results from the interviewees also show that these pupils have different ways of learning, like everybody else, and therefore the learning situations have to be varying. Finally, the result shows that the pupils want to have challenges, work together in groups and more support from teachers to be able to develop their full mathematical capacity.

The aim of this study was, through the eyes of six students from Socialhögskolan in Lund, to analyze how the students related to whiteness as a norm and their experiences of the norm in Swedish social work. How did they reflect upon its impact on the relationship between social workers and clients? Through semi-structured interviews with students, both within and outside the norm, the perspective was given of both expected discrimination and experienced discrimination due to color of skin and family-name. The tools which were used to analyze the empirics where postcolonial theory and social constructivism, theories which both question the “natural order” of society and the norms we all live by. An important question in this study is thus how whiteness as a norm exists and reproduces as a social construction. This study shows the complexity of the concept “whiteness as a norm” and that the construction of it works both ways. The stories conducted by the students showed existing discriminatory structures affecting the students on a personal level. They did also witness how people at their workplace were treated differently due to physical appearance and/or family name.

Biodiversity declines in farmland have been attributed to intensification of farming at the field level and loss of heterogeneity at the landscape level. However, farmers are not solely optimizing production; their actions are also influenced by social factors, tradition and interest in nature, which indirectly influence biodiversity but rarely are incorporated in studies of farmland biodiversity. We used social science methods to quantify farmers’ interest in nature on 16 farms with winter wheat fields in central Sweden, and combined this with biodiversity inventories of five organism groups (weeds, carabid beetles, bumblebees, solitary bees, and birds) and estimates of landscape composition andmanagement intensity at the field level.Agricultural intensity,measured as crop density, and farmers’ interest in nature explained variation in biodiversity, measured as the proportion of the regional species richness found on single fields. Interest in nature seemed to incorporate many actions taken by farmers and appeared to be influenced by both physical factors, for example, the surrounding landscape, and social factors, for example, social motivations.This study indicates that conservation of biodiversity in farmland, and design of new agri-environmental subsidy systems, would profit from taking farmers’ interest in nature and its relation to agricultural practices into account.

Plan recognition addresses the problem of inferring an agents goals from its action. Applications range from anticipating care-takers' needs to predicting volatile situations. In this contribution, we describe a prototype plan recognition system that is based on the well-researched theory of (weighted) finite tree automata. To illustrate the system?s capabilities, we use data gathered from matches in the real-time strategy game StarCraft II. Finally, we discuss how more advanced plan operators can be accommodated for in this framework while retaining computational efficiency by taking after the field of formal model checking and over-approximating the target language.

This study analyzed examples of sustainable ecosystem-based agriculture where management methods supported livelihoods of smallholders while at the same time local ecosystem services were enhanced in Ethiopia, Brazil, and the Philippines. Participation by farmers and collective actions were found to be a crucial driving force, as local specific knowledge and learning by doing were main components of the development. Social cohesion, particularly through associations and cooperatives, and improved marketing opportunities were also important drivers. Furthermore, recognition by authorities at all levels was perceived as crucial. Effects of climate change, insecure property rights, and political instability were potential threats. The possibilities of such systems to be scaled up beyond self-sufficiency raised further questions., AuthorCount:7

Today’s intensive agriculture needs to be transformed into sustainable production of food, and this process requires good tools that can assess whether an action is leading towards this in the long term. A critical issue is what optimal yield comprises in terms of other functions of agriculture, as higher yields might lead to e.g. a reduction in biodiversity or soil carbon. In this study, emergy analysis and footprinting were combined to assess and illustrate the total resource use caused by a farming activity (milk production) and to identify the renewable fraction of this resource use. The total efficiency was defined as a function of the resource use and the multifunctionality of production. The classification of ecosystem services in the Millennium Ecosystem Assessment (MA) was used as the basis for ranking multifunctionality. The results were expressed in the form of ecosystem bundles for the four MA categories (provisioning, supporting, regulating and cultural functions). Three scenarios with different degrees of input intensity and milk production were constructed and compared with the current production mode. The ratio of local renewable resource use to total resource use differed greatly between the different production strategies, being 1:3 for a self-sufficient organic farm and 1:14 for a conventional farm with maximum milk yield. Milk production was five-fold higher on the conventional farm, while generation of ecosystem services increased with increasing self-sufficiency under the local conditions prevailing in the study. Ecosystem services in all categories except provisioning were ranked higher when self-sufficiency increased.

This paper addresses the question of whether local selling of farm products improves on-farm biodiversity in rural areas. In contrast to the main agricultural trend of farms specializing and increasing in size in response to national and global markets, increasing numbers of Swedish farmers are diverting their efforts towards selling at local markets. Based on case studies of six farms selling their products locally, this paper explores the nature of the diversity on these farms and identifies qualities in the interaction between the farmers and their consumers that are supporting this diversity. The study showed that farmers who interacted with consumers were encouraged to diversify their production. Marketing a large diversity of products at a local market led to better income for participating farmers. Animal farms maintained important biodiversity associated with their extensive way of rearing animals on semi-natural pastures. Access to local markets promoted this., QC 20110131

Lantbrukaren är en nyckelperson i arbetet med odlingslandskapets biologiska mångfald. Kunskap om värdefulla arter och miljöer på den egna gården är avgörande i detta arbete. Vi har undersökt hur man kan arbeta med biologisk mångfald på gårdsnivå och samtidigt förmedla värdet av denna mångfald till konsumenter och samhälle. I det deltagardrivna projektet medverkade lantbrukare, rådgivare och forskare.Inventeringsmallar användes för att kartlägga den arter och miljöer. Genomgångarna på varandras gårdar och diskussionerna om tänkbara åtgärder hjälpte oss att ta fram en modell för hur man kan arbeta med biologisk mångfald. Det viktigaste är att det finns en variation i åtgärder och en anpassning till lokala förutsättningar. Många gånger kan små och kostnadseffektiva åtgärder vara väl så värdefulla för mångfalden.

QC 20120525

We show that a previously proposed algorithm for the N-best trees problem – not to be confused with the easier N-best runs problem – can be made more efficient by changing how it arranges and explores the search space. Given an integer N and a weighted tree automaton (wta) M over the tropical semiring, the algorithm computes N trees of minimal weight with respect to M. Compared to the original algorithm, the modifications increase the laziness of the evaluation strategy, which makes the new algorithm asymptotically more efficient than its predecessor.

Graph-based semantic representations are valuable in natural language processing, where it is often simple and effective to represent linguistic concepts as nodes, and relations as edges between them. Several attempts has been made to find a generative device that is sufficiently powerful to represent languages of semantic graphs, while at the same allowing efficient parsing. We add to this line of work by introducing graph extension grammar, which consists of an algebra over graphs together with a regular tree grammar that generates expressions over the operations of the algebra. Due to the design of the operations, these grammars can generate graphs with non-structural reentrancies, a type of node-sharing that is excessively common in formalisms such as abstract meaning representation, but for which existing devices offer little support. We provide a parsing algorithm for graph extension grammars, which is proved to be correct and run in polynomial time.

The N-best extraction problem consists in selecting the N highest ranking hypotheses from a set of hypotheses, with respect to a given ranking system. In our setting, the hypotheses and ranking are jointly represented by a weighted tree automaton (wta) over the tropical semiring: the hypotheses are trees, or runs on trees, and the ranking is decided by the weight assigned to them. In previous work, we presented an algorithm for N-best extraction that combines techniques to restrict the search space, and proved it to be correct and efficient. The algorithm is now implemented in the software Betty, allowing us to complement the deductive study with an empirical investigation. In particular, we compare our algorithm to the state-of-the-art algorithm for extracting the N best runs, implemented in in the software toolkit Tiburon. The data sets used in the experiments are wtas resulting from real-world natural language processing tasks, as well as artificially created wtas with varying degrees of nondeterminism. We find that Betty outperforms Tiburon on all tested data sets.