Abstract:Deep neural networks (DNNs) are being widely applied for various real-world applications across domains due to their high performance (e.g., high accuracy on image classification). Nevertheless, a well-trained DNN after deployment could oftentimes raise errors during practical use in the operational environment due to the mismatching between distributions of the training dataset and the potential unknown noise factors in the operational environment, e.g., weather, blur, noise etc. Hence, it poses a rather important problem for the DNNs' real-world applications: how to repair the deployed DNNs for correcting the failure samples (i.e., incorrect prediction) under the deployed operational environment while not harming their capability of handling normal or clean data. The number of failure samples we can collect in practice, caused by the noise factors in the operational environment, is often limited. Therefore, It is rather challenging how to repair more similar failures based on the limited failure samples we can collect. In this paper, we propose a style-guided data augmentation for repairing DNN in the operational environment. We propose a style transfer method to learn and introduce the unknown failure patterns within the failure data into the training data via data augmentation. Moreover, we further propose the clustering-based failure data generation for much more effective style-guided data augmentation. We conduct a large-scale evaluation with fifteen degradation factors that may happen in the real world and compare with four state-of-the-art data augmentation methods and two DNN repairing methods, demonstrating that our method can significantly enhance the deployed DNNs on the corrupted data in the operational environment, and with even better accuracy on clean datasets.

Repairs and Breaks Prediction for Deep Neural Networks

Interpretability Based Neural Network Repair

Technical Briefing on Deep Neural Network Repair

DeepRepair: Style-guided repairing for deep neural networks in the real-world operational environment.(2021)

DeepRepair: Style-Guided Repairing for Deep Neural Networks in the Real-World Operational Environment

Predicting the Number of Software Faults using Deep Learning

RNNrepair: Automatic RNN repair via model-based analysis

Repairing Neural Networks by Leaving the Right Past Behind

Repairing Deep Neural Networks: Fix Patterns and Challenges

Deep Learning for Just-In-Time Defect Prediction

DeepRepair: Style-Guided Repairing for DNNs in the Real-world Operational Environment

An Automated Monitoring and Repairing System for DNN Training

A deep neural network algorithm-based approach for predicting recovery period of accidents according to construction scale

MetaRepair: Learning to Repair Deep Neural Networks from Repairing Experiences

Software Defect Prediction via Deep Belief Network

NNrepair: Constraint-based Repair of Neural Network Classifiers

<i>DeepPatch</i>: A Patching-Based Method for Repairing Deep Neural Networks

Use of Deep Learning Model with Attention Mechanism for Software Fault Prediction

Repairing Neural Networks for Safety in Robotic Systems using Predictive Models

Disrupting Downtime: Different Deep Learning Journeys into Predictive Maintenance Anomaly Detection

Repairing Deep Neural Networks Based on Behavior Imitation