Institute of Information Science Academia Sinica
講 題: WALA (Watson Libraries for Analysis) Everywhere
講 者: Julian Dolby 先生 (IBM's Thomas J. Watson Research Center )
時 間: 2018-12-11 (Tue) 10:00 – 12:00
地 點: 資訊所新館106演講廳
邀請人: 陳郁方

Computation has spread to a dizzying array of devices---from supercomputers to desktops to laptops to phones to wristwatches to cars to home appliances---and taken on an extraordinary range of tasks---from climate modeling to flying aeroplanes to medical diagnosis to language translation to travel directions to movie recommendations.  This has increased the diversity of platforms on which software is developed and run; every new task seems to emphasize diverse languages, e.g. JavaScript for the Web, tools, e.g. Jupyter for Python, and platforms, e.g. Android for mobile.  To be useful, software development tools must support this diversity.  I will discuss how the Watson Libraries for Analysis (WALA) does so by supporting a range of runtime platforms,  programming languages, and development environments.  The last part will show how WALA components are supporting this diversity of languages and platforms for other analysis frameworks.  I shall present the range that WALA supports and show concrete examples in each case, some with demos.


Julian Dolby has been a Research Staff Member at IBM's Thomas J. Watson Research Center since 2000. He works on a range of topics, including static program analysis, software testing, the semantic web (AI) and programming technology support for machine learning.

- Initial exploration has revealed challenges with common machine learn practice that necessitate techniques like the following: always make every tensor dimension be a different size, to minimize matrix manipulation bugs; another tactic is copious comments detailing the layout of tensors.  They made some initial efforts on using program analysis to obviate such burdens and make code more reliable; they have started building such support using WALA:

- His program analysis work has recently been focused on scripting languages like JavaScript and on security analysis of Web applications; their work has been included in IBM products, most notably Rational AppScan products, and he is one of the primary authors of the publicly-available Watson Libraries for Analysis (WALA) program analysis infrastructure.  A presentation of their JavaScript work at

- His testing work has been focused on several areas: Web applications in the Apollo project, finding concurrency bugs using both dynamic execution and model checking, and finding security issues in Android apps (

- His semantic Web work has been on scalable inference with the SHER project; recently, he has focused on representing RDF data efficiently in an RDBMS.  A summary of much of their Semantic Web work in keynote he gave at the Semantic Big Data workshop at SIGMOD 2017 can be found here:

He was educated at the University of Wisconsin-Madison as an undergraduate, and at the University of Illinois at Urbana-Champaign as a graduate student where he worked with Professor Andrew Chien on programming systems for massively-parallel machines.