Made available in DSpace on 2016-04-04T18:36:24Z (GMT). No. of bitstreams: 1 Sergio Furgeri.pdf: 942403 bytes, checksum: a07b12b6b1cb4ccb6059ade6feece09f (MD5) Previous issue date: 2006-10-18 / The Information Science studies new. forms of information and knowledge representation for improving efficiency of information retrieval. The dissertation focus on information and knowledge representation, investigate the match and differential points between documentary languages, available in Information Science and markup languages, developed and used in Computer Science, with the objective of identify actions, theories and necessary processes for a bigger integration between the two areas. For this, it is established a revision of the necessary basic elements to the representation of the information and knowledge in the scope of the Information Science. To become possible the comparison between areas, consecrated knowledge and information representation models available in Information Science are presented, such as metadates, thesaurus and ontologies. In the scope of the Internet, the techniques of representation with the use of markup languages more used are presented and its contributions for the development of the Web Semantics. It is locked in presenting a proposal of structure of representation for information resources, especially developed for the Internet. The proposal was developed from the existing resources in the Computer Science, particularly, its available in language XML. It contemplates the definition of a ontology and culminates with the creation of a structure in XML to store metadates of electronic articles. / A Ci?ncia da Informa??o vem estudando formas de representa??o da informa??o e do conhecimento visando a recupera??o da informa??o. Esta pesquisa tem seu foco na representa??o do conhecimento e da informa??o, procurando investigar quais s?o os pontos convergentes e divergentes entre as linguagens document?rias da Ci?ncia da Informa??o e as linguagens de marca??o desenvolvidas e utilizadas na Ci?ncia da Computa??o, tendo em vista identificar as a??es, teorias e processos necess?rios para uma maior integra??o entre as duas ?reas. Para isso, faz-se uma revis?o dos elementos fundamentais necess?rios a represental??o da informa??o e do conhecimento no ?mbito da Ci?ncia da Informa??o. Para tornar poss?vel a compara??o entre ?reas, apresentam-se os modelos mais consagrados de representa??o do conhecimento e da informa??o provenientes da Ci?ncia da Informa??o, tais como metadados, tesauros e ontologias. No ?mbito da Internet, apresentam-se as t?cnicas de representa??o com o uso das linguagens de marca??o mais utilizadas e suas contribui??es para o desenvolvimento da Web Sem?ntica. Encerra-se apresentando uma proposta de estrutura de representa??o para recursos informacionais, especialmente os disponibilizados pela Internet. A proposta foi desenvolvida a partir dos recursos existentes na Ci?ncia da Computa??o, particularmente, os oferecidos pela linguagem XML. Contempla a defini??o de uma ontologia e culmina com a cria??o de uma estrutura em XML para armazenar metadados de artigos eletr?nicos.

When issuing surveys, having the option for free-text answer fields is only feasible where the number of respondents is small, as the work to summarize the answers becomes unmanageable with a large number of responses. Using NLP techniques to cluster these answers and summarize them would allow a greater range of survey creators to incorporate free-text answers in their survey, without making their workload too large. Academic work in this domain is sparse, especially for smaller languages such as Swedish. The Swedish company iMatrics is regularly hired to do this kind of summarizing, specifically for workplace-related surveys. Their method of clustering has been semiautomatic, where both manual preprocessing and postprocessing have been necessary to accomplish this task. This thesis aims to explore if using more advanced, unsupervised NLP text representation methods, namely SentenceBERT and Sent2Vec, can improve upon these results and reduce the manual work needed for this task. Specifically, three questions are to be answered. Firstly, do the methods show good results? Secondly, can they remove the time-consuming postprocessing step of combining a large number of clusters into a smaller number? Lastly, can a model where unsupervised learning metrics can be shown to correlate to the real-world usability of the model, thus indicating that these metrics can be used to optimize the model for new data? To answer these questions, several models are trained, employed, and then compared using both internal and external metrics: Sent2Vec, SentenceBERT, and traditional baseline models. A manual evaluation procedure is performed to assess the real-world usability of the clusterings looks like, to see how well the models perform as well as to see if there is any correlation between this result and the internal metrics for the clustering. The results indicate that improving the text representation step is not sufficient for fully automating this task. Some of the models show promise in the results of human evaluation, but given the unsupervised nature of the problem and the large variance between models, it is difficult to predict the performance of new data. Thus, the models can serve as an improvement to the workflow, but the need for manual work remains.

This thesis addresses key issues of scientific realism in the philosophy of biology and chemistry through investigation of an underexplored research domain: olfaction theory, or the science of smell. It also provides the first systematic overview of the development of olfactory practices and research into the molecular basis of odours across the 19th and 20th century. Historical and contemporary explanations and modelling techniques for understanding the material basis of odours are analysed with a specific focus on the entrenchment of technological process, research tradition and the definitions of materiality for understanding scientific advancement. The thesis seeks to make sense of the explanatory and problem solving strategies, different ways of reasoning and the construction of facts by drawing attention to the role and application of scientific representations in olfactory practices. Scientific representations such as models, classifications, maps, diagrams, lists etc. serve a variety of purposes that range from the stipulation of relevant properties and correlations of the research materials and the systematic formation of research questions, to the design of experiments that explore or test particular hypotheses. By examining a variety of modelling strategies in olfactory research, I elaborate on how I understand the relation between representations and the world and why this relation requires a pluralist perspective on scientific models, methods and practices. Through this work I will show how a plurality of representations does not pose a problem for realism about scientific entities and their theoretical contexts but, on the contrary, that this plurality serves as the most reliable grounding for a realistic interpretation of scientific representations of the world and the entities it contains. The thesis concludes that scientific judgement has to be understood through its disciplinary trajectory, and that scientific pluralism is a direct consequence of the historicity of scientific development.