Multimodal Software Analysis and Code Search

Multimodal Software Analysis and Code Search

时间： 地点：九龙湖校区计算机楼二楼会议室

报告简介：

Software systems become more and more complex and ubiquitous, integrating into many aspects of human life. Program analysis has been recognized as an important kind of techniques to improve software usability and reliability with increased developer productivity. However, we realize that program code does not exist by itself. There are many "contexts" that shape the structure and functionality of a piece of code. Such contexts include but are not limited to surrounding relevant pieces of code for a given piece of code, the logs of developer actions that create and change the piece of code, the traces and feedbacks of users that record the uses of the code, and environment conditions under which the code is deployed. The diverse "contexts" imply that program analysis on code itself may not be enough; often, we need to broaden our view and take various kinds of contexts as input to enable a more encompassing multimodal software analysis. This talk will give a broad overview of our work in software analysis using various kinds of data besides code, in particular sample applications in code search, and call for further exploration and collaborations on the rich opportunities in this direction. The first example utilizes developer feedbacks to improve the accuracy of code search that can return better code fragments for developers to reuse. The second example utilizes data available in the context of increasingly popular "social programming" on various project hosting websites to identify similar applications for the purpose of quick prototyping and code reuse for developers. The third example discusses a software project search engine with a domain specific language supporting the search of diverse software artifacts. Further examples and potentials of using multimodal software analysis will be introduced, and interested researchers and students are more than welcome to discuss and work together further.

报告人简介：

蒋凌霄，生于浙江，现任新加坡管理大学信息系统学院助理教授。分别于2000年和2003年获取北京大学数学学科学院信息科学学士学位和应用数学硕士学位。2009年获取加利福尼亚大学戴维斯分校计算机科学博士学位。2009年曾任美国英伟达公司软件测试策略师。曾在各类国际相关领域会议上发表四十多篇论文，包括多篇软件工程领域顶级会议（ICSE和FSE）论文。主要从事程序分析技术的研究和应用，包括软件测试、维护、安全、重用、软件库挖掘等，开发了多种快速、高效的重复代码探测、软件自动测试、检错算法和工具，致力于研究开发有助于保证软件质量、提高开发效率、降低维护成本的方法和技术。更多相关研究信息可联系 lxjiang@smu.edu.sg，或浏览: http://www.mysmu.edu/faculty/lxjiang/ 新加坡管理大学信息系统学院成立于2003年，近几年飞速发展。作为世界iSchools组织的50多个成员学院之一，我们注重信息和信息技术在实际商务环境中的用户、使用和影响。我们从事很多技术领域的研究和应用，例如软件工程、移动和普适系统、数据挖掘、人工智能、计算机安全等等。我们也从事信息管理领域的研究和应用，以及技术和管理的交叉领域。我们的博士课程每年招收10到20名全职或者交换研究生，给予全额奖学金，优秀者有机会赴本学院的合作大学卡内基梅隆大学参加为期10个月的培训。2015秋季入学申请已开始，截止日期2015年1月31日。更多相关信息可联系 lxjiang@smu.edu.sg，或浏览: http://sis.smu.edu.sg/programmes/PhD/online-application

计算机网络和信息集成教育部重点实验室（东南大学）