Syntax-guided synthesis of Datalog programs

Si, Xujie; Lee, Woosuk; Zhang, Richard; Albarghouthi, Aws; Koutris, Paraschos; Naik, Mayur

doi:10.1145/3236024.3236034

Cited by 43 publications

(22 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the running example, we will have to enumerate more than 12 × 10 6 candidate programs before discovering the one shown in Equation 1. Unsurprisingly, most work on program synthesis has focused on reducing the size of this search space: in our context, tools such as ALPS and ProSynth restrict the search space by only looking for programs composed of rules from a fixed finite set of candidate rules [47,52], while ILASP constrains the space through łmode declarationsž that bound the number of joins (in our case conjunctions) and the number of variables used [31,32]. On the other hand, Scythe, a synthesis tool for SQL queries, first finds łabstractž queries that over-approximate the desired output, and then searches for concrete instantiations of these abstract queries that are consistent with the data [57].…”

Section: Broadwaymentioning

confidence: 99%

Example-guided synthesis of relational queries

Thakkar

Naik

Sands

et al. 2021

Proceedings of the 42nd ACM SIGPLAN International Conference on Programming Language Design and Implementation

Self Cite

View full text Add to dashboard Cite

Program synthesis tasks are commonly specified via inputoutput examples. Existing enumerative techniques for such tasks are primarily guided by program syntax and only make indirect use of the examples. We identify a class of synthesis algorithms for programming-by-examples, which we call Example-Guided Synthesis (EGS), that exploits latent structure in the provided examples while generating candidate programs. We present an instance of EGS for the synthesis of relational queries and evaluate it on 86 tasks from three application domains: knowledge discovery, program analysis, and database querying. Our evaluation shows that EGS outperforms state-of-the-art synthesizers based on enumerative search, constraint solving, and hybrid techniques in terms of synthesis time, quality of synthesized programs, and ability to prove unrealizability. CCS Concepts:• Information systems → Relational database query languages; • Software and its engineering → Automatic programming; Programming by example; • Theory of computation → Constraint and logic programming.

show abstract

Section: Broadwaymentioning

confidence: 99%

Example-guided synthesis of relational queries

Thakkar

Naik

Sands

et al. 2021

Proceedings of the 42nd ACM SIGPLAN International Conference on Programming Language Design and Implementation

Self Cite

View full text Add to dashboard Cite

show abstract

“…Program synthesis commonly restricts the space of target concepts and biases the search to speed up computation and improve generalization. One form of bias has been to constrain the syntax: this has been formalized as the SyGuS problem [37] and as meta-rules in inductive logic programming [38], [39]. A meta-rule is construct of this form…”

Section: B the On-demand Feature Synthesis Algorithmmentioning

confidence: 99%

“…While SyGuS has been used in various applications [70]- [80], none of them aims to synthesize a provably sound static analyzer from data. While some of these existing techniques can synthesize Datalog rules [39], [81], [82], the focus has been on efficiency, e.g., pruning the search space based on syntactic structures, instead of guaranteeing the soundness of the analyzer. Power Side-Channel Analysis.…”

Section: Related Workmentioning

confidence: 99%

Data-Driven Synthesis of Provably Sound Side Channel Analyses

Wang

Sung

Raghothaman

et al. 2021

Preprint

View full text Add to dashboard Cite

We propose a data-driven method for synthesizing static analyses to detect side-channel information leaks in cryptographic software. Compared to the conventional way of manually crafting such static analyzers, which can be tedious, error prone and suboptimal, our learning-based technique is not only automated but also provably sound. Our analyzer consists of a set of type-inference rules learned from the training data, i.e., example code snippets annotated with the ground truth. Internally, we use syntax-guided synthesis (SyGuS) to generate new recursive features and decision tree learning (DTL) to generate analysis rules based on these features. We guarantee soundness by proving each learned analysis rule via a technique called query containment checking. We have implemented our technique in the LLVM compiler and used it to detect power side channels in C programs that implement cryptographic protocols. Our results show that, in addition to being automated and provably sound during synthesis, our analyzer can achieve the same empirical accuracy as two state-of-the-art, manually-crafted analyzers while being 300X and 900X faster, respectively.• We propose the first data-driven method for learning a provably sound static analyzer using syntax guided synthesis (SyGuS) and decision tree learning (DTL). • We guarantee soundness by formulating and solving a Datalog query containment checking problem. • We demonstrate the effectiveness of our method for detecting side channels in cryptographic software. In the remainder of this paper, we begin by presenting the technical background in Section II and our motivating example in Section III. We then describe the learner in Section IV and the prover in Section V, followed by the experimental results in Section VI. Finally, we survey the related work in Section VII and conclude in Section VIII. II. PRELIMINARIES A. Power Side-ChannelsPrior works in side-channel security [19]- [21] show that variance in the power consumption of a computing device may leak secret information; for example, when a secret value is stored in a physical register, its number of logical-1 bits may affect the power consumption of the CPU. Such side-channel leaks are typically mitigated by masking, e.g., using d random bits (r 1 , . . . , r d ) to split a key bit into d + 1 secret shares: key 1 = r 1 , . . ., key d = r d , and key d+1 = r 1 ⊕r 2 . . .⊕r d ⊕key, where ⊕ denotes the logical operation exclusive or (XOR). Since all d + 1 shares are uniformly distributed in the {0, 1}, in theory, this order-d masking scheme is secure in that any combination of less than d shares cannot reveal the secret, but combining all d + 1 shares, key 1 ⊕ key 2 ⊕ ...key d+1 = key, recovers the secret.In practice, masking countermeasures must also be implemented properly to avoid de-randomizing any of the secret shares accidentally. ConsiderWhile syntactically dependent on the two randomized values t L and t R , t is in fact leaky because, semantically, it does not depend on the random input r 1 . In this work, we aim t...

show abstract

“…However, this technique imposes an upper bound on the number of clauses and atoms in the Datalog program. The Alps tool [42] also performs Datalog program synthesis from examples but additionally requires meta-rule templates. In contrast, our technique focuses on a recursion-free subset of Datalog, but it does not require additional user input beyond examples and learns from failed synthesis attempts by using the concept of minimal distinguishing projections.…”

Section: Related Workmentioning

confidence: 99%

“…Our work is related to inductive logic programming (ILP) where the goal is to synthesize a logic program consistent with a set of examples [34,35,40,27,31,50,19,41]. Among ILP techniques, our work is most similar to recent work on Datalog program synthesis [3,42]. In particular, Zaatar [3] encodes an under-approximation of Datalog semantics using the theory of arrays and reduces synthesis to SMT solving.…”

Section: Related Workmentioning

confidence: 99%

Data migration using datalog program synthesis

et al. 2020

View full text Add to dashboard Cite

This paper presents a new technique for migrating data between different schemas. Our method expresses the schema mapping as a Datalog program and automatically synthesizes a Datalog program from simple input-output examples to perform data migration. This approach can transform data between different types of schemas (e.g., relational-tograph, document-to-relational) and performs synthesis efficiently by leveraging the semantics of Datalog. We implement the proposed technique as a tool called Dynamite and show its effectiveness by evaluating Dynamite on 28 realistic data migration scenarios.

show abstract

Syntax-guided synthesis of Datalog programs

Cited by 43 publications

References 46 publications

Example-guided synthesis of relational queries

Example-guided synthesis of relational queries

Data-Driven Synthesis of Provably Sound Side Channel Analyses

Data migration using datalog program synthesis

Contact Info

Product

Resources

About