 |
Gail Kaiser is a Professor of Computer Science in the Computer Science Department at Columbia University, where she is the Director of the Programming Systems Lab (PSL) and affiliated with the Software Systems Lab (SSL). Her interests include program analysis, software testing, software security, software systems, AI for software engineering (AI4SE) and software engineering for AI (SE4AI). Prof. Kaiser conducts research in software engineering and security from a systems perspective, focusing on program analysis and software testing. In the 1980s and early 1990s, Kaiser investigated semantics-focused extensions to language-based editors and process-oriented team software development environments, forerunners to today's IDEs and Continuous Integration, and in the mid 1990s through early 2000s she investigated collaborative work technologies leveraging the nascent World Wide Web and self-adaptation for the then-emerging cloud computing, particularly techniques for retrofitting legacy software. Beginning with her sabbatical at Columbia's Center for Computational Learning Systems in 2005-2006, Kaiser was among the first to investigate software engineering testing techniques, such as metamorphic testing, for finding bugs in machine learning software. Her more recent work ranges across static and dynamic program analysis techniques for both source code and binaries. She currently investigates secure computing paradigms and machine learning techniques for solving software engineering problems. She is greatly entertained to recently re-discover her early paper titled AI Techniques in Software Engineering, an invited chapter in this book published in 1990. The 1989 tech report version is here. The first paragraph says "The idea of using artificial intelligence techniques to support programming has been around for a long time. The earliest notion was to avoid programming entirely. The human user would just tell the computer what to do, without saying how to do it, and the computer would do the right thing. Even if this were feasible, however, it would be much too tedious, since each time the user would have to repeat the details of what he wanted done. So the goal of programming was to explain things to the computer only once, and then later on be able to tell the computer to do the same thing again in some short form, such as the name of the "program." Thus the idea evolved that a user would somehow tell the computer what program was desired, and the computer would write down the program in some internal form so that it could be remembered and repeated later. The assumption was that the resulting program would be correct, complete, efficient, easy to use, and so forth. It would also be exactly what the human user wanted." She wrote this in 1989, without any help from ChatGPT! |
