Abstract:Deep neural networks have experimentally demonstrated superior performance over other machine learning approaches in decision-making predictions. However, one major concern is the closed set nature of the classification decision on the trained classes, which can have serious consequences in safety critical systems. When the deep neural network is in a streaming environment, fast interpretation of this classification is required to determine if the classification result is trusted. Un-trusted classifications can occur when the input data to the deep neural network changes over time. One type of change that can occur is concept evolution, where a new class is introduced that the deep neural network was not trained on. In the majority of deep neural network architectures, the only option is to assign this instance to one of the classes it was trained on, which would be incorrect. The aim of this research is to detect the arrival of a new class in the stream. Existing work on interpreting deep neural networks often focuses on neuron activations to provide visual interpretation and feature extraction. Our novel approach, coined DeepStreamCE, uses streaming approaches for real-time concept evolution detection in deep neural networks. DeepStreamCE applies neuron activation reduction using an autoencoder and MCOD stream-based clustering in the offline phase. Both outputs are used in the online phase to analyse the neuron activations in the evolving stream in order to detect concept evolution occurrence in real time. We evaluate DeepStreamCE by training VGG16 convolutional neural networks on combinations of data from the CIFAR-10 dataset, holding out some classes to be used as concept evolution. For comparison, we apply the data and VGG16 networks to an open-set deep network solution - OpenMax. DeepStreamCE outperforms OpenMax when identifying concept evolution for our datasets.

Discovering an Evolutionary Classifier over a High-speed Nonstatic Stream

New algorithm for online classification over data streams based on max-frequency patterns

A Framework for On-Demand Classification of Evolving Data Streams

Online learning for streaming data classification in nonstationary environments

An Ensemble Classifier Algorithm for Mining Data Streams Based on Concept Drift

A Low-Granularity Classifier for Data Streams with Concept Drifts and Biased Class Distribution

On Demand Classification of Data Streams

Adaptive regularized ensemble for evolving data stream classification

Anticipative Bayesian classification for data streams with verification latency

Transfer estimation of evolving class priors in data stream classification

Feature Selection in the Data Stream Based on Incremental Markov Boundary Learning

DeepStreamCE: A Streaming Approach to Concept Evolution Detection in Deep Neural Networks

Online Boosting Adaptive Learning under Concept Drift for Multistream Classification

An Ensemble Classifier Method for Classifying Data Streams with Recurrent Concept Drift.

A new online learning algorithm for streaming data and decision support with a Bayesian approach

A grid density based framework for classifying streaming data in the presence of concept drift

Online Active Learning for Drifting Data Streams

A comprehensive ensemble classification techniques detecting and managing concept drift in dynamic imbalanced data streams

Evolutionary Multi-Objective Optimisation for Fairness-Aware Self Adjusting Memory Classifiers in Data Streams

A reliable adaptive prototype-based learning for evolving data streams with limited labels

Streaming Active Deep Forest for Evolving Data Stream Classification