Data Mining Practical Machine Learning Tools and Techniques

tical perspective. For instance, there are only three passing references to kuru in a book of 650 pages. This edition reflects the renewed interest in the immunological theories of multiple sclerosis. More than half the text is devoted to Professor Lumsden's analysis of the pathoIogy and, in particular, the chemical pathology of the immune response. There is a great deal of original work devoted to the chemistry and behaviour of the immunoglobulins. Much of this appears in specialist journals and one must be grateful for the critical summary provided here. Professor Lumsden unequivocally sees the key to the problem of multiple sclerosis in the study of its immunochemistry, relegating infection by a virus or a slow virus to a quite subsidiary role. The clinical studies drawing on wide practical experience help to get one's prejudices about the illness onto a more reasoned footing. The section on treatment is still sadly limited. Dr. McAlpine found little to add to the regime which he described in 1955. McAlpine, Lumsden, and Acheson's reappraisal is an essential reference for the practising neurologist and the new edition makes important modification of and changes in emphasis from the edition of 1965.

https://jnnp.bmj.com/content/jnnp/36/5/894.3.full.pdf

The assessment of aphasia and related disorders

Distributed word representations, or word vectors, have recently been applied to many tasks in natural language processing, leading to state-of-the-art performance. A key ingredient to the successful application of these representations is to train them on very large corpora, and use these pre-trained models in downstream tasks. In this paper, we describe how we trained such high quality word representations for 157 languages. We used two sources of data to train these models: the free online encyclopedia Wikipedia and data from the common crawl project. We also introduce three new word analogy datasets to evaluate these word vectors, for French, Hindi and Polish. Finally, we evaluate our pre-trained word vectors on 10 languages for which evaluation datasets exists, showing very strong performance compared to previous models.

Learning Word Vectors for 157 Languages

Distributed word representations have recently been proven to be an invaluable resource for NLP. These representations are normally learned using neural networks and capture syntactic and semantic information about words. Information about word morphology and shape is normally ignored when learning word representations. However, for tasks like part-of-speech tagging, intra-word information is extremely useful, specially when dealing with morphologically rich languages. In this paper, we propose a deep neural network that learns character-level representation of words and associate them with usual word representations to perform POS tagging. Using the proposed approach, while avoiding the use of any handcrafted feature, we produce state-of-the-art POS taggers for two languages: English, with 97.32% accuracy on the Penn Treebank WSJ corpus; and Portuguese, with 97.47% accuracy on the Mac-Morpho corpus, where the latter represents an error reduction of 12.2% on the best previous known result.

/pdf/learning-character-level-representations-for-part-of-speech-i2dwfr91wo.pdf

Learning Character-level Representations for Part-of-Speech Tagging

Most recent sentence simplification systems use basic machine translation models to learn lexical and syntactic paraphrases from a manually simplified parallel corpus. These methods are limited by the quality and quantity of manually simplified corpora, which are expensive to build. In this paper, we conduct an in-depth adaptation of statistical machine translation to perform text simplification, taking advantage of large-scale paraphrases learned from bilingual texts and a small amount of manual simplifications with multiple references. Our work is the first to design automatic metrics that are effective for tuning and evaluating simplification systems, which will facilitate iterative development for this task.

/pdf/optimizing-statistical-machine-translation-for-text-2uqxl8rf9e.pdf

Optimizing Statistical Machine Translation for Text Simplification

Facilita: auxílio à leitura de textos disponíveis na web

Com o intuito de avaliar a submissao de consultas em lingua natural, especificamente em portugues, a maquinas de busca na Web, e contrastar com as consultas por palavras-chave, realizou-se um experimento com alunos, professores e funcionarios de uma universidade brasileira. Particularmente, analisaram-se as consultas para verificar se os usuarios expressavam bem seus objetivos em palavras-chave; como expressariam seus objetivos em lingua natural, caso esta possibilidade fosse oferecida; se as consultas em lingua natural forneciam informacoes que pudessem facilitar a recuperacao de informacao. O pedido de colaboracao foi enviado a 440 pessoas de um instituto de computacao da universidade. Foram obtidas 63 consultas, correspondentes a 42 objetivos. Observou-se que, para o item a, na maioria dos casos (71,43%), as consultas por meio de palavras-chave nao trazem todas as informacoes declaradas importantes no objetivo; para o item b as consultas foram feitas por meio de perguntas (71,87%), afirmacoes (18,75%) e ordens (9,37%); e, para o item c todas as perguntas diretas deixavam claro o objetivo da consulta ja com a primeira palavra da frase, ou com as duas ou tres primeiras, com excecao das iniciadas pela palavra "qual".

http://revista.ibict.br/ciinf/article/download/1014/1069

Como incrementar a qualidade dos resultados das máquinas de busca: da análise de logs à interação em português

We report on an ongoing project aimed at developing software tools to help non-native users of English in the writing of scientific papers. The environment, named AMADEUS, (AMiable Article Development Environment for User Support), is based on an empirical resource which consists in providing linguistic input to the user in the form of expressions and templates stored in a sentences base, and is at an initial exploratory stage, in the sense that we are trying different user-interfaces and environments for developing the tools. One of the versions conceived for the environment is already implemented.

Writing tools for non-native users of english

Semantic Role Labeling (SRL) is a Natural Language Processing task that enables the detection of events described in sentences and the participants of these events. For Brazilian Portuguese (BP), there are two studies recently concluded that perform SRL in journalistic texts. [1] obtained F1-measure scores of 79.6, using the PropBank.Br corpus, which has syntactic trees manually revised; [8], without using a treebank for training, obtained F1-measure scores of 68.0 for the same corpus. However, the use of manually revised syntactic trees for this task does not represent a real scenario of application. The goal of this paper is to evaluate the performance of SRL on revised and non-revised syntactic trees using a larger and balanced corpus of BP journalistic texts. First, we have shown that [1]’s system also performs better than [8]’s system on the larger corpus. Second, the SRL system trained on non-revised syntactic trees performs better over non-revised trees than a system trained on gold-standard data.

Automatic Semantic Role Labeling on Non-revised Syntactic Trees of Journalistic Texts

In this paper we introduce the AMADEUS strategy, which has been used to produce scientific writing tools for non-native users of English for 15 years, and emphasize a learn-by-doing approach through which students and novice writers can improve their scientific writing. More specifically, we provide a 9-step recipe for the students to compile writing material according to a procedure that has proven efficient in scientific writing courses.

Sandra Maria Aluísio

Papers

Facilita: auxílio à leitura de textos disponíveis na web

Como incrementar a qualidade dos resultados das máquinas de busca: da análise de logs à interação em português

Writing tools for non-native users of english

Automatic Semantic Role Labeling on Non-revised Syntactic Trees of Journalistic Texts

Developing strategies to produce better scientific papers: a Recipe for non-native users of English