Retrieving and classifying instances of source code plagiarism

Automatic detection of source code plagiarism is an important research field for both the commercial software industry and within the research community. Existing methods of plagiarism detection primarily involve exhaustive pairwise document comparison, which does not scale well for large software collections. To achieve scalability, we approach the problem from an information retrieval (IR) perspective. We retrieve a ranked list of candidate documents in response to a pseudo-query representation constructed from each source code document in the collection. The challenge in source code document retrieval is that the standard bag-of-words (BoW) representation model for such documents is likely to result in many false positives being retrieved, because of the use of identical programming language specific constructs and keywords. To address this problem, we make use of an abstract syntax tree (AST) representation of the source code documents. While the IR approach is efficient, it is essentially unsupervised in nature. To further improve its effectiveness, we apply a supervised classifier (pre-trained with features extracted from sample plagiarized source code pairs) on the top ranked retrieved documents. We report experiments on the SOCO-2014 dataset comprising 12K Java source files with almost 1M lines of code. Our experiments confirm that the AST based approach produces significantly better retrieval effectiveness than a standard BoW representation, i.e., the AST based approach is able to identify a higher number of plagiarized source code documents at top ranks in response to a query source code document. The supervised classifier, trained on features extracted from sample plagiarized source code pairs, is shown to effectively filter and thus further improve the ranked list of retrieved candidate plagiarized documents.     

More than Bags of Words: Sentiment Analysis with Word Embeddings

ABSTRACT

Moving beyond the dominant bag-of-words approach to sentiment analysis we introduce an alternative procedure based on distributed word embeddings. The strength of word embeddings is the ability to capture similarities in word meaning. We use word embeddings as part of a supervised machine learning procedure which estimates levels of negativity in parliamentary speeches. The procedure’s accuracy is evaluated with crowdcoded training sentences; its external validity through a study of patterns of negativity in Austrian parliamentary speeches. The results show the potential of the word embeddings approach for sentiment analysis in the social sciences.     

Effects of interval aerobic training combined with strength exercise on body composition,glycaemic and lipid profile and aerobic capacity of obese rats

The purpose of this study was to investigate the effects of interval aerobic training combined with strength exercise in the same training session on body composition, and glycaemic and lipid profile in obese rats. Sixteen lean Zucker rats and sixteen obese Zucker rats were randomly divided into exercise and sedentary subgroups (4 groups, n = 8). Exercise consisted of interval aerobic training combined with strength exercise in the same training session. The animals trained 60 min/day, 5 days/week for 8 weeks. Body composition, lipid and glycaemic profiles and inflammatory markers were assessed.

Results showed that fat mass was reduced in both lean and obese rats following the exercise training (effect size (95% confidence interval (CI)) = 1.8 (0.5–3.0)). Plasma low-density lipoprotein–cholesterol and fasting glucose were lower in the exercise compared to the sedentary groups (d = 2.0 (0.7–3.2) and 1.8 (0.5–3.0), respectively). Plasma insulin was reduced in exercise compared to sedentary groups (d = 2.1 (0.8–3.4)). Some exercise × phenotype interactions showed that the highest decreases in insulin, homeostatic model assessment-insulin resistance, fasting and postprandial glucose were observed in the obese + exercise group (all, P < 0.01). The findings of this study suggest that interval aerobic training combined with strength exercise would improve body composition, and lipid and glycaemic profiles, especially in obese rats.     

Many choices,one destination: multimodal university brand construction in an urban public transportation system

Amidst global competition in higher education, colleges and universities adopt strategies that mimic and adapt business practices. Branding is now a widespread practice in higher education; multimodal advertisement is a manifestation of emerging branding strategies for universities. While the visibility of brands in higher education has grown substantially in recent years, its empirical study has lagged. This article reports on the findings from a study that employed social semiotic and multimodal analysis to explore text and visual rhetoric as brand construction strategies in publicly displayed university advertisements. After analyzing photographs representing advertising campaigns from 16 different universities displayed in an urban public transportation system, I suggest that universities construct their brand identity through messages that emphasize multiple choices and convenience, but construct success primarily according to corporate standards and values.     

Equity and social exclusion measures in EU lifelong learning policies

ABSTRACT

The aim of this article is to examine how EU lifelong learning policies are trying to reach the vulnerable by looking at what measures against social exclusion they offer and how equitable these measures are. It is a qualitative study that focuses on policy documents that form the European Union’s legal and political frameworks of reference in the lifelong learning area since 1992. The document analysis has been complemented by semi-structured interviews with EU lifelong learning experts. The findings show that early school leavers and migrants are the main target groups in the policies, leaving many other groups at risk of being excluded from learning opportunities. There is not enough attention to measures addressing wider social phenomena. There is also an overemphasis on basic skills which are understood in a very narrow way as literacy and numeracy when referring to the vulnerable. We argue that a greater variety of measures as well as better targeted measures are needed to address the multiple and complex needs of the vulnerable. Such measures would allow a broader understanding of lifelong learning where those that are hardest to reach are offered learning opportunities independent of their personal and social circumstances     

Pedro Gutiérrez Bueno’s Textbooks: Audiences, Teaching Practices and Chemical Revolution

Pedro Gutiérrez Bueno wrote two editions of a chemistry textbook between 1788 and 1802. The paper offers a comparative view of both editions taking into account Gutiérrez Bueno’s biography, his intended audience and the changes related to the so-called chemical revolution. Some conclusions are at odds with common images about scientific revolutions and the role of textbooks in science. The analysis aims to shed some light on the multiple historical forces shaping textbooks and the way in which scientific change is assimilated by their writers. This research was partially supported by the program BHA2002-04611-CO3-02