Learning to Represent Programs with Property Signatures

By Augustus Odena, Charles Sutton. 2020

34 citations recorded by the source corpus.

We introduce the notion of property signatures, a representation for programs and program specifications meant for consumption by machine learning algorithms. Given a function with input type $\tau_{in}$ and output type $\tau_{out}$, a property is a function of type: $(\tau_{in}, \tau_{out}) \rightarrow \texttt{Bool}$ that (informally) describes some simple property of the function under consideration. For instance, if $\tau_{in}$ and $\tau_{out}$ are both lists of the same type, one property might ask `is the input list the same length as the output list?'. If we have a list of such properties, we can evaluate them all for our function to get a list of outputs that we will call the property signature. Crucially, we can `guess' the property signature for a function given only a set of input/output pairs meant to specify that function. We discuss several potential applications of property signatures and show experimentally that they can be used to improve over a baseline synthesizer so that it emits twice as many programs in less than one-tenth of the time.

Subfields: Artificial Intelligence, Information Systems, Software, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Quoc V. Le
2. Charles Sutton
3. Luke S. Zettlemoyer
4. Percy Liang
5. Augustus Odena
6. Graham Neubig
7. Mohit Bansal
8. Denny Zhou

In Information Systems

1. Carrie J. Cai
2. Augustus Odena
3. Jacob Austin
4. Maarten Bosma
5. David Dohan
6. Henryk Michalewski
7. Maxwell Nye
8. Quoc V. Le

In Software

1. Augustus Odena
2. Henryk Michalewski
3. Maarten Bosma
4. David Dohan
5. Jacob Austin
6. Quoc V. Le
7. Michael Terry
8. Carrie J. Cai

In Natural Language Processing

1. Charles Sutton
2. Augustus Odena
3. Quoc V. Le
4. David Dohan
5. Henryk Michalewski
6. Maxwell Nye
7. Maarten Bosma
8. Jacob Austin

Citations in the corpus

• Program Synthesis with Large Language Models (2021, methodology): “, 2020], but these techniques have largely been applied to restricted domain-specific languages (DSLs) [Gulwani, 2011] or to languages that are more fully featured but that nevertheless are designed specifically with synthesis in mind [Odena and Sutton, 2020].”
• MISIM: An End-to-End Neural Code Similarity System (2020, background): “In recent years, there has been an emergence of many MP systems, due, in part, to advances in machine learning, formal methods, data availability, and computing effi-ciency [1, 3–5, 10, 13, 29, 32, 41, 44].”