DeepGauge: Multi-Granularity Testing Criteria for Deep Learning Systems

By Lei Ma, Felix Juefei-Xu, Fuyuan Zhang, Jiyuan Sun, Minhui Xue, Bo Li, Chunyang Chen, Ting Su, Li Li, Yang Liu and 2 more. International Conference on Automated Software Engineering, 2018

708 citations recorded by the source corpus.

Deep learning (DL) defines a new data-driven programming paradigm that constructs the internal system logic of a crafted neuron network through a set of training data. We have seen wide adoption of DL in many safety-critical scenarios. However, a plethora of studies have shown that the state-of-the-art DL systems suffer from various vulnerabilities which can lead to severe consequences when applied to real-world applications. Currently, the testing adequacy of a DL system is usually measured by the accuracy of test data. Considering the limitation of accessible high quality test data, good accuracy performance on test data can hardly provide confidence to the testing adequacy and generality of DL systems. Unlike traditional software systems that have clear and controllable logic and functionality, the lack of interpretability in a DL system makes system analysis and defect detection difficult, which could potentially hinder its real-world deployment. In this paper, we propose DeepGauge, a set of multi-granularity testing criteria for DL systems, which aims at rendering a multi-faceted portrayal of the testbed. The in-depth evaluation of our proposed testing criteria is demonstrated on two well-known datasets, five DL systems, and with four state-of-the-art adversarial attack techniques against DL. The potential usefulness of DeepGauge sheds light on the construction of more generic and robust DL systems.

Subfields: Artificial Intelligence, Software, Signal Processing

View Paper: https://www.semanticscholar.org/paper/de9194d974530e816ee9b33e224fe7c955915fb1

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

In Software

In Signal Processing

Citations in the corpus

DeepHunter: a coverage-guided fuzz testing framework for deep neural networks (2019, methodology): “For example, the neuron coverage [36], k -multisection coverage [31], etc .”
Provable repair of deep neural networks (2021, positive): “More recent research tackles testing of DNNs [23, 42, 43, 45, 47, 54, 56, 58].”
FakeSpotter: A Simple Baseline for Spotting AI-Synthesized Fake Faces (2019, background): “Neuron coverage is widely adopted in assuring the quality and security of DNNs and plays an important role in analyzing DNNs [25,26,27,28,29].”
Bayes-Probe: Distribution-Guided Sampling for Prediction Level Sets (2020, background): “Several other neural network testing paradigms have been proposed (Wicker et al., 2018; Pei et al., 2017; Ma et al., 2018).”
Superion: Grammar-Aware Greybox Fuzzing (2018, neutral): “, compiler fuzzing [17, 19, 38, 39, 43, 60], kernel fuzzing [18, 32, 57], IoT (Internet of Things) fuzzing [15], OS fuzzing [48], smart contract fuzzing [36], GUI testing [59], and deep learning system testing [44].”
Machine Learning Testing: Survey, Landscapes and Horizons (2019, methodology): “Their work pioneered to propose coverage criteria for DNN. Enlightened by this paper, a number of machine learning testing techniques have emerged, such as DeepTest [76], DeepGauge [92], DeepConcolic [111], and DeepRoad [79].”
Stealthy and Efficient Adversarial Attacks against Deep Reinforcement Learning (2020, background): “Recently, how to systematically generate adversarial inputs were studied in DNN testing scenarios (Ma et al. 2018; Guo et al. 2019; Xie et al. 2019a; 2019b; Du et al. 2019).”
Model Agnostic Defence Against Backdoor Attacks in Machine Learning (2019, background): “A recent work DeepGauge [14] measures the test quality and formalises a set of testing criteria for DNNs. SafeCV [30] takes a feature guided black-box approach to verify the safety of DNNs. DeepConcolic [20] performs concolic testing to discover violations of robustness.”
Trustworthy AI: From Principles to Practices (2021, methodology): “Such coverage criteria are effective for certifying the robustness of a DNN against adversarial attacks [226].”
DeepMutation: Mutation Testing of Deep Learning Systems (2018, methodology): “Currently, testing for DL software is still at an early stage, with some initial research work focused on accuracy and neuron coverage, such as DeepXplore [12], DeepGauge [13],”