Maria Pittou

Maria Pittou · Mon, 04.01.2016


Our research involves the simplification and modification of computer aided controllers’ synthesis problem. Since most of the existing work results in bulky controllers, we aim at developing techniques for deriving controllers which are efficient but simple to understand. The ExCAPE Summer School is an up to date event which is strongly related with our field of research. ExCAPE targets on the theoretical and practical advances of software synthesis, i.e., the development of software that exploits the foundations of program synthesis. In particular, the main topics of the second ExCAPE Summer School, that we participated in, included reactive synthesis, syntax-guided synthesis, and inductive learning.

In particular, throughout the tutorial "Reactive Synthesis" given by Prof. R. Bloem, we had the opportunity to improve our knowledge on the fundamental elements and methods of Reactive Synthesis which is in the core of our Ph.D. The tutorial was quite interactive, we were given the chance to discuss on the presented material, and frequently we were asked to answer questions. Most importantly, we had the chance to experience a recently available tool, the so called RATSY, which can be used to synthesize reactive systems from their temporal specification (expressed in GR(1)), and among its features it is also included a game-based approach for debugging specifications. It is important to mention that we had no previous experience on RATSY, and although such tools are available online, working with them under the supervision of one of the developers of it, allowed us for face to face conversation, and thus of a better understanding.
 
Another interesting tutorial was "Syntax-Guided Synthesis" given by Prof. A. Solar-Lezama. Syntax-Guided Synthesis is a distinct recent approach of synthesis, where a partial system is obtained from the given incomplete details, and then the synthesizer fills in the missing details using user-specified assertions as the correctness specification. The difference from the common approaches of synthesis consists in the fact that in addition to the correctness specification, a syntactic template for the desired system is also provided. Throughout the tutorial, we became familiar with this synthesis problem, and with the relevant tools (SyGuS tool, CEGIS tool for counterexamples in guided synthesis). We emphasize on this topic because it belongs to the recent trends of the field that enable the user to define and restrict the space of the allowed implementations. Obviously this is related with our research since we target on efficient methods for obtaining simpler implementations (controllers), and guided synthesis is simpler in terms of requiring incomplete information and efficient since it allows for pruning out candidate implementations. Since we can potentially consider such approaches, or their variations and improvements for our framework, we consider their knowledge a necessary asset for our work.
All the topics presented during the ExCAPE Summer School are more or less related with our research interests, and therefore their attendance has been interesting and useful. However, due to space limitations we mentioned above only couple of them.
 
Generally, it is essential that we keep up to date with the latest advances in our field, hence participating at ExCAPE helped towards this direction. Apart from this, we had the opportunity to be in a live face to face setting with colleagues who are active in our line of work, and with professors who are leaders of the field. All in all, attending the ExCAPE Summer School has been a unique experience that assisted our work.