English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Programming-by-example is the task of synthesizing a program that isconsistent with a set of user-provided input-output examples. As examples areoften an under-specification of one's intent, a good synthesizer must choosethe intended program from the many that are consistent with the given set ofexamples. Prior work frames program synthesis as a cooperative game between alistener (that synthesizes programs) and a speaker (a user choosing examples),and shows that models of computational pragmatic inference are effective inchoosing the user intended programs. However, these models requirecounterfactual reasoning over a large set of programs and examples, which isinfeasible in realistic program spaces. In this paper, we propose a novel wayto amortize this search with neural networks. We sample pairs of programs andexamples via self-play between listener and speaker models, and use pragmaticinference to choose informative training examples from this sample.We then usethe informative dataset to train models to improve the synthesizer's ability todisambiguate user-provided examples without human supervision. We validate ourmethod on the challenging task of synthesizing regular expressions from examplestrings, and find that our method (1) outperforms models trained withoutchoosing pragmatic examples by 23% (a 51% relative increase) (2) matches theperformance of supervised learning on a dataset of pragmatic examples providedby humans, despite using no human data in training.

Generating Pragmatic Examples to Train Neural Program Synthesizers