Code completion with statistical language models

Abstract

We address the problem of synthesizing code completions for programs using APIs. Given a program with holes, we synthesize completions for holes with the most likely sequences of method calls.Our main idea is to reduce the problem of code completion to a natural-language processing problem of predicting probabilities of sentences. We design a simple and scalable static analysis that extracts sequences of method calls from a large codebase, and index these into a statistical language model. We then employ the language model to find the highest ranked sentences, and use them to synthesize a code completion. Our approach is able to synthesize sequences of calls across multiple objects together with their arguments.Experiments show that our approach is fast and effective. Virtually all computed completions typecheck, and the desired completion appears in the top 3 results in 90% of the cases.

Pythia: AI-assisted Code Completion System

“…A large number of intelligent code completion systems for both statically and dynamically typed languages have been proposed in the literature [4,5,2,6,7,8]. Best Matching Neighbor (BMN) and statistical language models such as n-grams, as well as recurrent neural network (RNN) based approaches leveraging sequential nature of the source code have been particularly effective at creating such systems.…”

On the Convergence of Artificial Intelligence and Distributed Ledger Technology: A Scoping Review and Future Research Agenda

“…Future research could expand this work by considering further classes of smart contracts [35] and further details, such as different smart contract compiler versions [36]. Furthermore, future research that aims to support software developers could aim at not only detecting security vulnerabilities in a smart contract, but also at localizing this vulnerability (e.g., by expressing the information which portions of the bytecode cause the vulnerabilities [37]) or even by providing suggestions to fix it [60,61].…”

Multi-token code suggestions using statistical language models

“…Other methods, such as that of Omar [4] and Raychev et al [5] have also focused on techniques that require more information about the language such as its syntax, or its type system.…”