Search

Taking the research area of Big Data as a case study, we propose an approach for exploring how academic topics shift through the interactions among audiences across different altmetric sources. Data used is obtained from Web of Science (WoS) and Altmetric.com, with a focus on Blog, News, Policy, Wikipedia, and Twitter. Author keywords from publications and terms from online events are extracted as the main topics of the publications and the online discussion of their audiences at Altmetric. Different measures are applied to determine the (dis)similarities between the topics put forward by the publication authors and those by the online audiences. Results show that overall there are substantial differences between the two sets of topics around Big Data scientific research. The main exception is Twitter, where high-frequency hashtags in tweets have a stronger concordance with the author keywords in publications. Among the online communities, Blogs and News show a strong similarity in the terms commonly used, while Policy documents and Wikipedia articles exhibit the strongest dissimilarity in considering and interpreting Big Data related research. Specifically, the audiences not only focus on more easy-to-understand academic topics related to social or general issues, but also extend them to a broader range of topics in their online discussions. This study lays the foundations for further investigations about the role of online audiences in the transformation of academic topics across altmetric sources, and the degree of concern and reception of scholarly contents by online communities.

We perform a risk assessment of the Public Safety Assessment (PSA), a software used in San Francisco and other jurisdictions to assist judges in deciding whether defendants need to be detained before their trial. With a mixed-methods approach including stakeholder interviews and the use of theoretical frameworks, we lay out the values at play as pretrial justice is automated. After identifying value implications of delegating decision making to technology, we articulate benefits and limitations of the PSA solution, as well as suggest mitigation strategies. We then draft the Handoff Tree, a novel algorithmic approach to pretrial justice that accommodates some of the inherent limitations of risk assessment tools by design. The model pairs every prediction with an associated error rate, and hands off the decision to the judge if the uncertainty is too high. By explicitly stating error rate, the Handoff Tree aims both to limit the impact of predictive disparity between race and gender, and to prompt judges to be more critical of retention recommendations, given the high rate of false positives they often entail., Comment: This report was presented at the Information Ethics Roundtable 2019 at Northeastern University, see https://cssh.northeastern.edu/philosophy/our-recent-past-talks/

Multiple studies have shown that gender balance in the fields of Science, Technology, Engineering and Maths -- and in particular in ICT -- is still far to be achieved. Several initiatives have been recently taken to increase the women participation, but it is difficult, at present, to evaluate their impact and their potential of changing the situation. This paper contributes to the discussion by presenting a descriptive analysis of the gender balance in Computer Science and Computer Engineering in Italian Universities.

Software engineers need to manage both technical and professional skills in order to be successful. Our university offers a 5.5 year program that mixes computer science, software and computer engineering, where the first two years are mostly math and physics courses. As such, our students' first real teamwork experience is during the introductory SE course, where they modify open source projects in groups of 6-8. However, students have problems working in such large teams, and feel that the course material and project are "disconnected". We decided to redesign this course in 2017, trying to achieve a balance between theory and practice, and technical and professional skills, with a maximum course workload of 150 hrs per semester. We share our experience in this paper, discussing the strategies we used to improve teamwork and help students learn new technologies in a more autonomous manner. We also discuss what we learned from the two times we taught the new course., 8 pages, 9 tables, 4 figures, Second International Workshop on Software Engineering Education for Millennials

Mobile phones are now widely adopted by most of the world population. Each time a call is made (or an SMS sent), a Call Detail Record (CDR) is generated by the telecom companies for billing purpose. These metadata provide information on when, how, from where and with whom we communicate. Conceptually, they can be described as a geospatial, dynamic, weighted and directed network. Applications of CDRs for development are numerous. They have been used to model the spread of infectious diseases, study road traffic, support electrification planning strategies or map socio-economic level of population. While massive, CDRs are not statistically representative of the whole population due to several sources of bias (market, usage, spatial and temporal resolution). Furthermore, mobile phone metadata are held by telecom companies. Consequently, their access is not necessarily straightforward and can seriously hamper any operational application. Finally, a trade-off exists between privacy and utility when using sensitive data like CDRs. New initiatives such as Open Algorithm might help to deal with these fundamental questions by allowing researchers to run algorithms on the data that remain safely stored behind the firewall of the providers., Comment: 28 pages, 8 figures, 1 table. Published in Ph.D. Thesis: Jacques, D.C. Harnessing the Data Revolution for Food Security and Poverty Mapping: synergies between Mobile Phone Data, Earth Observation and Official Statistics in Senegal. May 2018

In this paper, we design and develop a pricing model applicable to strategy proof pricing. To provide an economic stability towards its consumers. The economic model we use is Vickrey-Clarke-Groves (VCG). By this each service provider has to provide a true cost of its services in the cloud market. For the selection of suitable service for the consumer we adopt a dynamic programing based algorithm and VCG is used to calculate the payment. Strategy proof pricing offers a unique cloud pricing service that takes the complexity out of traditional pricing and enables cloud providers to price accurately, consistently and competitively, Comment: Includes 2 Figures, 2 Tables and 4 Pages. Presented in International Conference on Communication, Information and Computing Technology (ICCICT-15) held in Amritsar on 12-13 May, 2015

Cloud computing is an evolving utility computing mechanism in which cloud consumer can detect, choose and utilize the resources (infrastructure, software and platform) and provide service to user based on pay per use model as computing utilities. Current computing mechanism is effective, particular for medium and small cloud based companies, in which it permits easy and reliable access to cloud services like infrastructure, software and platform. Present cloud computing is almost similar to the existing models: cluster computing and grid computing. The important key technical features of cloud computing which includes autonomic service, rapid elasticity, end-to-end virtualization support, on-demand resource pooling and transparency in cloud billing. Further, non-technical features of cloud computing includes environment friendliness, little maintenance overhead, lower upfront costs, faster time to deployments, Service Level Agreement (SLA) and pay-as-you-go-model. In distributed computing environment, unpredictability of service is a fact, so same possible in cloud also. The success of next-generation Cloud Computing infrastructures will depend on how capably these infrastructures will discover and dynamically tolerate computing platforms, which meet randomly varying resource and service requirements of Cloud costumer applications., Comment: Author's Viewpoint on Autonomic Cloud Computing and Uploaded on Research Gate [https://www.researchgate.net/profile/Sukhpal_Gill]

The Dynamic Scalability of resources, a problem in Infrastructure as a Service (IaaS) has been the hotspot for research and industry communities. The heterogeneous and dynamic nature of the Cloud workloads depends on the Quality of Service (QoS) allocation of appropriate workloads to appropriate resources. A workload is an abstraction of work that instance or set of instances that are going to perform. Running a web service or being a Hadoop data node is valid workloads. The efficient management of dynamic nature resources can be done with the help of workloads. Until workload is considered a fundamental capability, the Cloud resources cannot be utilized in an efficient manner. In this paper, different workloads have been identified and categorized along with their characteristics and constraints. The metrics based on Quality of Service (QoS) requirements have been identified for each workload and have been analyzed for creating better application design., Including 5 Tables and 3 Figures, Presented in the International Conference on Next Generation Computing and Communication Technologies (ICNGCCT 2014), Dubai, UAE on 23-24 April, 2014

This article describes a formal strategy of geometric complexity theory (GCT) to resolve the {\em self referential paradox} in the $P$ vs. $NP$ and related problems. The strategy, called the {\em flip}, is to go for {\em explicit proofs} of these problems. By an explicit proof we mean a proof that constructs proof certificates of hardness that are easy to verify, construct and decode. The main result in this paper says that (1) any proof of the arithmetic implication of the $P$ vs. $NP$ conjecture is close to an explicit proof in the sense that it can be transformed into an explicit proof by proving in addition that arithmetic circuit identity testing can be derandomized in a blackbox fashion, and (2) stronger forms of these arithmetic hardness and derandomization conjectures together imply a polynomial time algorithm for a formidable explicit construction problem in algebraic geometry. This may explain why these conjectures, which look so elementary at the surface, have turned out to be so hard., 32 pages