Better reasoning about software engineering activities

2001

Abstract Machine learning is practical for software engineering problems, even in datastarved domains. When data is scarce, knowledge can be farmed from seeds; ie minimal and partial descriptions of a domain. These seeds can be grown into large datasets via Monte Carlo simulations. The datasets can then be harvested using machine learning techniques. Examples of this knowledge farming approach, and the associated technique of data-mining, is given from numerous software engineering domains.

Communications of the IIMA, 2006

Software Process Improvement (SPI) programs are frequently considered to be planned in nature. However, there is recent evidence to suggest that SPI can be understood as a form of learning. Drawing on the organizational learning literature this paper proposes an active learning perspective of improvements in processes. This view recognizes the various actors in the project to be reflective in their actions, making sense of the current context and thus designing their use of the process to best suit their needs at the time. The changes in the processes emerge through ongoing adjustments, experimentation and improvisation as developers and managers seek to improve their product development.

2005

Cognitive Informatics is a new research area that combines concepts from cognitive sciences and informatics. Particularly, classification according to cognitive styles or learning styles is a common practise in educational and business areas in order to obtain better performance from people involved in learning and collaborative tasks. Humanintensive processes in software engineering make it similar to those areas. In this paper, we propose using a kind of people classification to improve those processes of software engineering where human behaviour is a critical influence on their success.

Proceedings of the FSE/SDP workshop on Future of software engineering research - FoSER '10, 2010

Mining software engineering data has emerged as a successful research direction over the past decade. In this position paper, we advocate Software Intelligence (SI) as the future of mining software engineering data, within modern software engineering research, practice, and education. We coin the name SI as an inspiration from the Business Intelligence (BI) field, which offers concepts and techniques to improve business decision making by using fact-based support systems. Similarly, SI offers software practitioners (not just developers) up-to-date and pertinent information to support their daily decision-making processes. SI should support decisionmaking processes throughout the lifetime of a software system not just during its development phase.

Journal of Systems and Software, 1987

This paper reports on an empirical study of the cognitive processes involved in program comprehension. Verbal protocols were gathered from professional programmers as they were engaged in a program-understanding task. Based on analysis of these protocols, several types of interesting cognitive events were identified. These include asking questions and conjecturing facts about the code. We describe these event types and use them to derive a computational model of the programmers' mental processes.

Requirements Engineering, 1999

IEEE Transactions on Software Engineering, 1988

T h e maintenance of large information systems involves continuous modifications in response to evolving business conditions or changing user requirements. Based on evidence from a case study, we show that the systems maintenance activity would benefit greatly if the process knowledge reflecting the teleology of a design could be captured and used in order t o reason about the consequences of changing conditions or requirements. We describe a formalism called REMAP (REpresentation and MAintenance of Process knowledge) that accumulates design process knowledge t o manage systems evolution. T o accomplish this, REMAP acquires and maintains dependencies among the design decisions made during a prototyping process, and is able t o learn general domain-specific design rules on which such dependencies are based. This knowledge can not only be applied t o prototype refinement and systems maintenance, but can also support the re-use of existing design or software fragments t o construct similar ones using analogical reasoning techniques.

ACM SIGSOFT Software Engineering Notes, 2004

A program's behavior is ultimately the collection of all its executions. This collection is diverse, unpredictable, and generally unbounded. Thus it is especially suited to statistical analysis and machine learning techniques. The primary focus of this paper is on the automatic classification of program behavior using execution data. Prior work on classifiers for software engineering adopts a classical batchlearning approach. In contrast, we explore an active-learning paradigm for behavior classification. In active learning, the classifier is trained incrementally on a series of labeled data elements. Secondly, we explore the thesis that certain features of program behavior are stochastic processes that exhibit the Markov property, and that the resultant Markov models of individual program executions can be automatically clustered into effective predictors of program behavior. We present a technique that models program executions as Markov models, and a clustering method for Markov models that aggregates multiple program executions into effective behavior classifiers. We evaluate an application of active learning to the efficient refinement of our classifiers by conducting three empirical studies that explore a scenario illustrating automated test plan augmentation.

2000

Building and evaluating prediction systems is an important activity for software engineering researchers. Increasing numbers of techniques and datasets are now being made available. Unfortunately systematic comparison is hindered by the use of different accuracy indicators and evaluation processes. We argue that these indicators are statistics that describe properties of the estimation errors or residuals and that the sensible choice of indicator is largely governed by the goals of the estimator. For this reason it may be helpful for researchers to provide a range of indicators. We also argue that it is useful to formally test for significant differences between competing prediction systems and note that where only a few cases are available this can be problematic, in other words the research instrument may have insufficient power. We demonstrate that this is the case for a well known empirical study of cost models. Simulation, however, could be one means of overcoming this difficulty.

2021

Automation and intelligence constitute a major preoccupation in the field of software engineering. With the great evolution of Artificial Intelligence, researchers and industry were steered to the use of Machine Learning and Deep Learning models to optimize tasks, automate pipelines, and build intelligent systems. The big capabilities of Artificial Intelligence make it possible to imitate and even outperform human intelligence in some cases as well as to automate manual tasks while rising accuracy, quality, and efficiency. In fact, accomplishing software-related tasks requires specific knowledge and skills. Thanks to the powerful capabilities of Artificial Intelligence, we could infer that expertise from historical experience using machine learning techniques. This would alleviate the burden on software specialists and allow them to focus on valuable tasks. In particular, Model-Driven Engineering is an evolving field that aims to raise the abstraction level of languages and to focus more on domain specificities. This allows shifting the effort put on the implementation and low-level programming to a higher point of view focused on design, architecture, and decision making. Thereby, this will increase the efficiency and productivity of creating applications. For its part, the design of metamodels is a substantial task in Model-Driven Engineering. Accordingly, it is important to maintain a high-level quality of metamodels because they constitute a primary and fundamental artifact. However, the bad design choices as well as the repetitive design modifications, due to the evolution of requirements, could deteriorate the quality of the metamodel. The accumulation of bad design choices and quality degradation could imply negative outcomes in the long term. Thus, refactoring metamodels is a very important task. It aims to improve and maintain good quality characteristics of metamodels such as maintainability, reusability, extendibility, etc. Moreover, the refactoring task of metamodels is complex, especially, when dealing with large designs. Therefore, automating and assisting architects in this task is advantageous since they could focus on more valuable tasks that require human intuition. In this thesis, we propose a cartography of the potential tasks that we could either automate or improve using Artificial Intelligence techniques. Then, we select the metamodeling vii task and we tackle the problem of metamodel refactoring. We suggest two different approaches: A first approach that consists of using a genetic algorithm to optimize set quality attributes and recommend candidate metamodel refactoring solutions. A second approach based on mathematical logic that consists of defining the specification of an input metamodel, encoding the quality attributes and the absence of smells as a set of constraints and finally satisfying these constraints using Alloy.

— Software is nowadays a critical component of our lives and everyday-work working activities. However, as the technological infrastructure of the modern world evolves a great challenge arises for developing high quality software systems with increasing size and complexity. Software engineers and researchers are striving to meet this challenge by developing and implementing software engineering methodologies able to deliver software products of high quality, within budget and time constraints. The field of machine learning in software engineering has recently emerged to provide means for addressing, studying, analyzing, and understanding critical software development issues and at the same time to offer mature machine learning techniques such as artificial neural network, Bayesian networks, decision trees, fuzzy logic, genetic algorithms, and rule induction. Machine learning algorithms have proven to be of great practical value to software engineering. Not surprisingly, the field of software engineering turns out to be a fertile ground where many software development tasks could be formulated as learning problems and approached in terms of learning algorithms. In this paper, we first take a look at the characteristics and applicability of some frequently utilized machine learning algorithms. We then present the application of machine learning in the different phases of software engineering that include project planning, requirements analysis, design, implementation, testing and maintenance.

Computer software and media applications, 2024

The integration of Artificial Intelligence (AI) into the space of software engineering marks a transformative period that reshapes traditional development processes and propels the industry into a new era of innovation. This exploration delves into the multifaceted impact of AI, from its roots in early symbolic AI to the contemporary dominance of machine learning and deep learning. AI's applications span various domains, but its significance in software engineering lies in its ability to enhance efficiency, improve software quality, and introduce novel approaches to problem-solving. From automating routine tasks to streamlining complex development workflows, AI acts as a virtual collaborator, allowing human developers to focus on higher-order thinking and creativity. This study introduces the application of AI in software engineering. reverse-engineering, and development environments. Moreover, ethical considerations, challenges, and future trends, including explainable AI, reinforcement learning, and human-AI collaboration, are presented.

2009

Product or company names used in this set are for identification purposes only. Inclusion of the names of the products or companies does not indicate a claim of ownership by IGI Global of the trademark or registered trademark. Library of Congress Cataloging-in-Publication Data Artificial intelligence applications for improved software engineering development : new prospects / Farid Meziane and Sunil Vadera, editors. p. cm. Includes bibliographical references and index.

2002

Software engineering (SE) truisms capture broadlyapplicable principles of software construction. The trouble with truisms is that such general principles may not apply in specific cases. This paper tests the specificity of two SE truisms: (a) increasing software process level is a desirable goal; and (b) it is best to remove errors during the early parts of a software lifecycle.