LARA

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
Next revision Both sides next revision
insynth [2013/05/17 19:00]
kuraj
insynth [2013/05/17 22:25]
vkuncak
Line 1: Line 1:
 ====== InSynth: Type-Driven Interactive Synthesis of Code Snippets ====== ====== InSynth: Type-Driven Interactive Synthesis of Code Snippets ======
  
-**InSynth** is a tool for interactive synthesis of code snippets. It solves quantiative version of the type inhabitation problem to synthesize ​a ranked list of code fragments that use given library functions.+**InSynth** is a tool for interactive synthesis of code snippets. It synthesizes ​a ranked list of code fragments that use given library functions.
  
-The tool is presented in the "​Complete Completion using Types and Weights"​ [[http://​lara.epfl.ch/​~kuncak/​papers/​GveroETAL13CompleteCompletionTypesWeights.pdf|paper]] published at PLDI'​2013 (technical report can be found [[http://​infoscience.epfl.ch/​record/​182807|here]]) and used in the framework for synthesis and verification of recursive functional programs described [[https://​infoscience.epfl.ch/​record/​186043?​ln=en|here]].+A recent version ​is presented in the following ​paper:
  
-Poster for "Complete Completion using Types and Weights" can be viewed ​[[https://​dl.dropboxusercontent.com/​u/​4182451/​insynth/​poster.pdf|here]].+ ​*[[http://​lara.epfl.ch/​~kuncak/​papers/​GveroETAL13CompleteCompletionTypesWeights.pdf|Complete Completion using Types and Weights]], PLDI'​2013  
 + 
 +[[https://​dl.dropboxusercontent.com/​u/​4182451/​insynth/​poster.pdf|Here is the corresponding poster]] 
 + 
 +These techniques are used in [[https://​infoscience.epfl.ch/​record/​186043?​ln=en|this framework for synthesis and verification of recursive functional programs]].
  
 ---- ----
Line 57: Line 61:
 === Training Data ===  === Training Data === 
 Training data we use to build the corpus can be found [[http://​laraserver.epfl.ch/​~gvero/​corpus.zip|here]]. Training data we use to build the corpus can be found [[http://​laraserver.epfl.ch/​~gvero/​corpus.zip|here]].
-