Infrared: A Meta Bug Detector

TL;DR

This paper introduces Meta Bug Detection (MBD), a novel learning-based bug detector that is bug-type generic, self-explainable, and sample-efficient, significantly improving bug detection across various bug types with less training data.

Contribution

The paper presents MBD, a new meta-learning approach for bug detection that overcomes limitations of existing methods by being bug-type generic, self-explainable, and requiring less data.

Findings

MBD effectively detects diverse bugs including null pointer dereference and data races.

MBD outperforms static analysis tools like Facebook Infer and anomaly detection methods like FICS.

MBD demonstrates high accuracy with reduced training data requirements.

Abstract

The recent breakthroughs in deep learning methods have sparked a wave of interest in learning-based bug detectors. Compared to the traditional static analysis tools, these bug detectors are directly learned from data, thus, easier to create. On the other hand, they are difficult to train, requiring a large amount of data which is not readily available. In this paper, we propose a new approach, called meta bug detection, which offers three crucial advantages over existing learning-based bug detectors: bug-type generic (i.e., capable of catching the types of bugs that are totally unobserved during training), self-explainable (i.e., capable of explaining its own prediction without any external interpretability methods) and sample efficient (i.e., requiring substantially less training data than standard bug detectors). Our extensive evaluation shows our meta bug detector (MBD) is effective in catching a variety of bugs including null pointer dereference, array index out-of-bound, file handle leak, and even data races in concurrent programs; in the process MBD also significantly outperforms several noteworthy baselines including Facebook Infer, a prominent static analysis tool, and FICS, the latest anomaly detection method.