Program synthesis using conflict-driven learning

Yu, Feng; Martins, Ruben; Bastani, Osbert; Dillig, Işıl

doi:10.1145/3192366.3192382

Cited by 91 publications

(97 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Automating table transformations. Our technique for synthesizing table transformations is related to several recent techniques for automating data wrangling [Feng et al 2018[Feng et al , 2017Harris and Gulwani 2011;Tran et al 2009;Wang et al 2017a;Zhang and Sun 2013]. Among these, Scythe generates SQL queries from input-output examples and prunes the search space by grouping partial queries into equivalence classes [Wang et al 2017a].…”

Section: Related Workmentioning

confidence: 99%

“…The Morpheus system automates table transformation tasks that arise in R programming and leverages logical specifications of R library functions to prune the search space using SMT-based reasoning [Feng et al 2017]. Morpheus's successor, Neo, generalizes this technique to other domains and further uses logical specifications to learn from failed synthesis attempts [Feng et al 2018]. A unifying theme among all these prior efforts is that the specification is a pair of concrete input and output tables.…”

Section: Related Workmentioning

confidence: 99%

“…Given the large search space that must be explored by program synthesizers, a common trick is to perform lightweight program analysis to prune the search space [Feser et al 2015;Polikarpova et al 2016;Wang et al 2017bWang et al , 2018b. The particular flavor of program analysis varies between different synthesizers and ranges from domain-specific deduction [Feser et al 2015;Wang et al 2017b] to abstract interpretation [Wang et al 2018b] to SMT-based reasoning [Feng et al 2018;Polikarpova et al 2016]. Furthermore, some of these techniques leverage program analysis to construct a compact version space [Polozov and Gulwani 2015;Wang et al 2018b] while others use it to prune partial programs in enumerative search [Feng et al 2018;Wang et al 2017b].…”

Section: Related Workmentioning

confidence: 99%

“…The particular flavor of program analysis varies between different synthesizers and ranges from domain-specific deduction [Feser et al 2015;Wang et al 2017b] to abstract interpretation [Wang et al 2018b] to SMT-based reasoning [Feng et al 2018;Polikarpova et al 2016]. Furthermore, some of these techniques leverage program analysis to construct a compact version space [Polozov and Gulwani 2015;Wang et al 2018b] while others use it to prune partial programs in enumerative search [Feng et al 2018;Wang et al 2017b]. Similar to these efforts, we also use program analysis to prove infeasibility of partial programs but with two key differences: First, our analysis is tailored to table transformation programs and infers inclusion constraints between tables.…”

Section: Related Workmentioning

confidence: 99%

See 3 more Smart Citations

Visualization by example

Wang

Bodík

et al. 2019

Proc. ACM Program. Lang.

Self Cite

View full text Add to dashboard Cite

While visualizations play a crucial role in gaining insights from data, generating useful visualizations from a complex dataset is far from an easy task. In particular, besides understanding the functionality provided by existing visualization libraries, generating the desired visualization also requires reshaping and aggregating the underlying data as well as composing different visual elements to achieve the intended visual narrative. This paper aims to simplify visualization tasks by automatically synthesizing the required program from simple visual sketches provided by the user. Specifically, given an input data set and a visual sketch that demonstrates how to visualize a very small subset of this data, our technique automatically generates a program that can be used to visualize the entire data set.From a program synthesis perspective, automating visualization tasks poses several challenges that are not addressed by prior techniques. First, because many visualization tasks require data wrangling in addition to generating plots from a given table, we need to decompose the end-to-end synthesis task into two separate sub-problems. Second, because the intermediate specification that results from the decomposition is necessarily imprecise, this makes the data wrangling task particularly challenging in our context. In this paper, we address these problems by developing a new compositional visualization-by-example technique that (a) decomposes the end-to-end task into two different synthesis problems over different DSLs and (b) leverages bi-directional program analysis to deal with the complexity that arises from having an imprecise intermediate specification.We have implemented our visualization-by-example approach in a tool called Viser and evaluate it on 83 visualization tasks collected from on-line forums and tutorials. Viser can solve 84% of these benchmarks within a 600 second time limit, and, for those tasks that can be solved, the desired visualization is among the top-5 generated by Viser in 70% of the cases. building programming systems like D3 [Bostock et al. 2011] and Vega-Lite [Satyanarayan et al. 2017] to further facilitate real-world visualization tasks.Despite all these recent efforts, data visualization still remains a challenging task that requires considerable expertise -in fact, so much so that some companies even have job titles like "data visualization expert" or "data visualization specialist." Generally speaking, there are three key reasons that make data visualization a challenging task. First, beyond having a good insight about how the data can be best visualized, one needs to have sufficient knowledge about how to use the relevant visualization libraries. Second, different visualization primitives typically require the data to be in different formats; so, in order to experiment with different types of visualizations, one needs to constantly reshape the data into different formats. Finally, generating the intended visualization typically requires modifications to the original dataset, incl...

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Visualization by example

Wang

Bodík

et al. 2019

Proc. ACM Program. Lang.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Inductive programming has been successfully used for a variety of code synthesis tasks, most notably in the FlashFill approach to spreadsheet function generation [8]. In this work, a set of examples is provided by a user, and a sequence of inductive logic passes are applied to incrementally reduce the search space of possible programs which match the examples in a broadly similar way to SMT solving [4] [5]. This approach depends heavily on the use of a highly restricted and specialised language over which to search, often with the inductive logic passes being designed specifically with that language in mind.…”

Section: Introductionmentioning

confidence: 99%

General Program Synthesis Using Guided Corpus Generation and Automatic Refactoring

Wild

Porter

2019

Search-Based Software Engineering

View full text Add to dashboard Cite

Program synthesis aims to produce source code based on a user specification, raising the abstraction level of building systems and opening the potential for non-programmers to synthesise their own bespoke services. Both genetic programming (GP) and neural code synthesis have proposed a wide range of approaches to solving this problem, but both have limitations in generality and scope. We propose a hybrid search-based approach which combines (i) a genetic algorithm to autonomously generate a training corpus of programs centred around a set of highly abstracted hints describing interesting features; and (ii) a neural network which trains on this data and automatically refactors it towards a form which makes a more ideal use of the neural network's representational capacity. When given an unseen program represented as a small set of input and output examples, our neural network is used to generate a rank-ordered search space of what it sees as the most promising programs; we then iterate through this list up to a given maximum search depth. Our results show that this approach is able to find up to 60% of a human-useful target set of programs that it has never seen before, including applying a clip function to the values in an array to restrict them to a given maximum, and offsetting all values in an array.

show abstract

Testing for Race Conditions in Distributed Systems via SMT Solving

2020

View full text Add to dashboard Cite

Data races, a condition where two memory accesses to the same memory location occur concurrently, have been shown to be a major source of concurrency bugs in distributed systems. Unfortunately, data races are often triggered by non-deterministic event orderings that are hard to detect when testing complex distributed systems. In this paper, we propose Spider, an automated tool for identifying data races in distributed system traces. Spider encodes the causal relations between the events in the trace as a symbolic constraint model, which is then fed into an SMT solver to check for the presence of conflicting concurrent accesses. To reduce the constraint solving time, Spider employs a pruning technique aimed at removing redundant portions of the trace. Our experiments with multiple benchmarks show that Spider is effective in detecting data races in distributed executions in a practical amount of time, providing evidence of its usefulness as a testing tool.

show abstract

Program synthesis using conflict-driven learning

Cited by 91 publications

References 39 publications

Visualization by example

Visualization by example

General Program Synthesis Using Guided Corpus Generation and Automatic Refactoring

Testing for Race Conditions in Distributed Systems via SMT Solving

Contact Info

Product

Resources

About