CITATION.cff

# This CITATION.cff file was generated with cffinit.
# Visit https://bit.ly/cffinit to generate yours today!

cff-version: 1.2.0
title: >-
  Monitor-Guided Decoding of Code LMs with Static Analysis
  of Repository Context
message: >-
  If you use this repository, please cite it using the metadata
  from this file.
type: software
authors:
  - given-names: Lakshya A
    family-names: Agrawal
    email: t-lakagrawal@microsoft.com
    affiliation: Microsoft Research
    orcid: 'https://orcid.org/0000-0003-0409-8212'
  - given-names: Aditya
    family-names: Kanade
    email: kanadeaditya@microsoft.com
    affiliation: Microsoft Research
  - given-names: Navin
    family-names: Goyal
    email: navingo@microsoft.com
    affiliation: Microsoft Research
  - given-names: Shuvendu K.
    family-names: Lahiri
    email: shuvendu.lahiri@microsoft.com
    affiliation: Microsoft Research
  - given-names: Sriram K.
    family-names: Rajamani
    email: sriram@microsoft.com
    affiliation: Microsoft Research
identifiers:
  - type: doi
    value: 10.48550/arXiv.2306.10763
  - type: url
    value: >-
      https://openreview.net/forum?id=qPUbKxKvXq&noteId=98Ukj82fSP
abstract: >-
  Language models of code (LMs) work well when the
  surrounding code provides sufficient context. This is not
  true when it becomes necessary to use types, functionality
  or APIs defined elsewhere in the repository or a linked
  library, especially those not seen during training. LMs
  suffer from limited awareness of such global context and
  end up hallucinating.


  Integrated development environments (IDEs) assist
  developers in understanding repository context using
  static analysis. We extend this assistance, enjoyed by
  developers, to LMs. We propose monitor-guided decoding
  (MGD) where a monitor uses static analysis to guide the
  decoding. We construct a repository-level dataset
  PragmaticCode for method-completion in Java and evaluate
  MGD on it. On models of varying parameter scale, by
  monitoring for type-consistent object dereferences, MGD
  consistently improves compilation rates and agreement with
  ground truth. Further, LMs with fewer parameters, when
  augmented with MGD, can outperform larger LMs. With MGD,
  SantaCoder-1.1B achieves better compilation rate and
  next-identifier match than the much larger
  text-davinci-003 model.


  We also conduct a generalizability study to evaluate the
  ability of MGD to generalize to multiple programming
  languages (Java, C# and Rust), coding scenarios (e.g.,
  correct number of arguments to method calls), and to
  enforce richer semantic constraints (e.g., stateful API
  protocols). Our data and implementation are available at
  https://github.com/microsoft/monitors4codegen.
keywords:
  - program analysis
  - correctness
  - code generation
  - Language models