Showing 1–2 of 2 results for author: Feser, J K

Deductive Optimization of Relational Data Storage

Authors: John K. Feser, Samuel Madden, Nan Tang, Armando Solar-Lezama

Abstract: Optimizing the physical data storage and retrieval of data are two key database management problems. In this paper, we propose a language that can express a wide range of physical database layouts, going well beyond the row- and column-based methods that are widely used in database management systems. We use deductive synthesis to turn a high-level relational representation of a database query int… ▽ More Optimizing the physical data storage and retrieval of data are two key database management problems. In this paper, we propose a language that can express a wide range of physical database layouts, going well beyond the row- and column-based methods that are widely used in database management systems. We use deductive synthesis to turn a high-level relational representation of a database query into a highly optimized low-level implementation which operates on a specialized layout of the dataset. We build a compiler for this language and conduct experiments using a popular database benchmark, which shows that the performance of these specialized queries is competitive with a state-of-the-art in memory compiled database system. △ Less

Submitted 5 February, 2020; v1 submitted 7 March, 2019; originally announced March 2019.

Differentiable Functional Program Interpreters

Authors: John K. Feser, Marc Brockschmidt, Alexander L. Gaunt, Daniel Tarlow

Abstract: Programming by Example (PBE) is the task of inducing computer programs from input-output examples. It can be seen as a type of machine learning where the hypothesis space is the set of legal programs in some programming language. Recent work on differentiable interpreters relaxes the discrete space of programs into a continuous space so that search over programs can be performed using gradient-bas… ▽ More Programming by Example (PBE) is the task of inducing computer programs from input-output examples. It can be seen as a type of machine learning where the hypothesis space is the set of legal programs in some programming language. Recent work on differentiable interpreters relaxes the discrete space of programs into a continuous space so that search over programs can be performed using gradient-based optimization. While conceptually powerful, so far differentiable interpreter-based program synthesis has only been capable of solving very simple problems. In this work, we study modeling choices that arise when constructing a differentiable programming language and their impact on the success of synthesis. The main motivation for the modeling choices comes from functional programming: we study the effect of memory allocation schemes, immutable data, type systems, and built-in control-flow structures. Empirically we show that incorporating functional programming ideas into differentiable programming languages allows us to learn much more complex programs than is possible with existing differentiable languages. △ Less

Submitted 2 March, 2017; v1 submitted 7 November, 2016; originally announced November 2016.