Dyna: Extending Datalog for Modern AI

Abstract

Abstract. Modern statistical AI systems are quite large and complex; this interferes with research, development, and education. We point out that most of the computation involves database-like queries and updates on complex views of the data. Specifically, recursive queries look up and aggregate relevant or potentially relevant values. If the results of these queries are memoized for reuse, the memos may need to be updated through change propagation. We propose a declarative language, which generalizes Datalog, to support this work in a generic way. Through examples, we show that a broad spectrum of AI algorithms can be concisely captured by writing down systems of equations in our notation. Many strategies could be used to actually solve those systems. Our examples motivate certain extensions to Datalog, which are connected to functional and object-oriented programming paradigms. Why a New Data-Oriented Language for AI?Modern AI systems are frustratingly big, making them time-consuming to engineer and difficult to modify. In this chapter, we describe our work toward a declarative language that was motivated originally by various use cases in AI. Our goal is to make it easier to specify a wide range of new systems that are more or less in the mold of existing AI systems. Our declarative language should simplify inferential computation in the same way that the declarative language of regular expressions has simplified string pattern matching and transduction.All areas of AI have become data-intensive, owing to the flood of data and the pervasiveness of statistical modeling and machine learning. A system's extensional data (inputs) include not only current sensory input but also background ⋆ This chapter has been condensed for publication; the full version is available as [22].

LogicBlox, Platform and Language: A Tutorial

“…For instance, the following is a fragment of a program that solves Sudoku puzzles via linear programming: In the example, X is the existentially-quantified predicate predicate, and Obj is the objective function for the solver. (The last rule uses an aggregate syntax, +=, inspired by the Dyna project [8]. )…”

Efficient verification of railway infrastructure designs against standard regulations

“…This method is used in the Semantic Web tool RDFox, 11 which has a high performance on multicore processors with in-memory databases. We are considering RDFox as an alternative candidate for the back end of our incremental railway infrastructure verification procedure.…”

“…It claims support for incremental verification, but we could not evaluate it on our railway example due to absence of freely downloadable distributions. Dyna is a promising new Datalog-like language for modern statistical AI systems [11]. It has currently not matured sufficiently for our application, but its techniques are promising, and we hope to see it more fully developed in the future.…”

kProbLog: an algebraic Prolog for machine learning

“…Today there exist many frameworks and formalisms that tightly integrate these two paradigms; they support probabilistic and logical inference as well as learning. Prominent examples include PRISM [52], Dyna [19,18], Markov Logic [49], BLOG [42], and ProbLog [12]. Statistical relational learning and probabilistic programming have enabled an entirely new generation of applications.…”

“…Standard probabilistic programming languages such as PRISM and ProbLog have clear and fixed semantics (the distribution semantics) that cannot be changed. These limitations have motivated the development of algebraic logical languages such as Dyna [19,18] and aProbLog [34]. While standard probabilistic programming languages such as PRISM and ProbLog label facts with probabilities, Dyna and aProbLog use algebraic labels belonging to a semiring, which allows the use of other algebraic structures than the probabilistic semi-ring on top of the underlying logic programs.…”