Time：9:00-11:00, October 21, 2020
Place: Rm. 2135, Science Building 2, Peking University
Online Conference ID：Tencent Meeting 279 831 622
Talk 1: Type-Based Resource-Guided Search
Speaker: Di Wang (Carnegie Mellon University)
Abstract: Resource usage—the amount of time, memory, and energy a program requires for its execution—is one of the central subjects of computer science. Recently, automatic amortized resource analysis (AARA) has been introduced as a type-based, compositional, and efficient approach for resource analysis. In this talk, I will talk about my research on integrating AARA with the search procedures for (i) compositional worst-case input generation, and (ii) type-directed synthesis with resource-bound guarantees. If time permits, I will also discuss ongoing work for extending AARA with probabilities.
Bio: Di Wang is a doctoral student in computer science at Carnegie Mellon University. His research areas are programming languages and software engineering, with a focus on probabilistic programming, type systems, static resource analysis, and program synthesis. Currently, he is working on language-level integrations for Bayesian inference and probabilistic programming systems.
Talk 2: Guiding Dynamic Programing via Structural Probability for Accelerating Programming by Example
Speaker: Ruyi Ji (Peking University)
Abstract: Programming by example (PBE) is an important subproblem of program synthesis, and PBE techniques have been applied to many domains. Though many techniques for accelerating PBE systems have been explored, the scalability remains one of the main challenges: There is still a gap between the performances of state-of-the-art synthesizers and the industrial requirement. To further speed up solving PBE tasks, in this paper, we propose a novel PBE framework MaxFlash. MaxFlash uses a model based on structural probability, named topdown prediction models, to guide a search based on dynamic programming, such that the search will focus on subproblems that form probable programs, and avoid improbable programs. Our evaluation shows that MaxFlash achieves 4.107 times - 2080 times speed-ups against state-of-the-art solvers on 244 real-world tasks.
Bio: Ruyi Ji is a Ph.D. student in computer science at Peking University. He is now studying program synthesis and program repair, advised by Professor Yingfei Xiong and Professor Zhenjiang Hu.