Abstract

Abstract
Runtime adaptability is expected to adjust the application and the mapping of computations according to usage contexts, operating environments, resources availability, etc. However, extending applications with adaptive features can be a complex task, especially due to the current lack of programming models and compiler support. One of the runtime adaptability possibilities is the use of specialized code according to data workloads and environments. Traditional approaches use multiple code versions generated offline and, during runtime, a strategy is responsible to select a code version. Moving code generation to runtime can achieve important improvements but may impose unacceptable overhead. This paper presents an aspect-oriented programming approach for runtime adaptability. We focus on a separation of concerns (strategies vs. application) promoted by a domain-specific language for programming runtime strategies. Our strategies allow runtime specialization based on contextual information. We use a template-based runtime code generation approach to achieve program specialization. We demonstrate our approach with examples from image processing, which depict the benefits of runtime specialization and illustrate how several factors need to be considered to eficiently adapt the application.

Abstract

Abstract
Since the advent of artificial intelligence, researchers have been trying to create machines that emulate human behaviour. Back in the 1960s however, Licklider (IRE Trans Hum Factors Electron 4-11, 1960) believed that machines and computers were just part of a scale in which computers were on one side and humans on the other (human computation). After almost a decade of active research into human computation and crowdsourcing, this paper presents a survey of crowdsourcing human computation systems, with the focus being on solving micro-tasks and complex tasks. An analysis of the current state of the art is performed from a technical standpoint, which includes a systematized description of the terminologies used by crowdsourcing platforms and the relationships between each term. Furthermore, the similarities between task-oriented crowdsourcing platforms are described and presented in a process diagram according to a proposed classification. Using this analysis as a stepping stone, this paper concludes with a discussion of challenges and possible future research directions.

Abstract

Abstract
The sheer size of data sets from application domains such as biomedical and social networks will lead to the need to develop algorithms that have strict time bounds and can tolerate temporary unavailability of data if they are to produce acceptable results in feasible time. In this paper we describe a simple, yet powerful, object-based concurrent programming model that features atomicity, timed execution and tolerance to data unavailability. We describe the underlying concepts and illustrate their use in a sample computation on large graphs. This experience shows that it is possible to augment existing concurrent programming models to support developers in developing and reasoning about incomplete computations, we believe, will become, an increasingly important class of algorithms. © 2015 IEEE.