To cope with real-world dynamics, an intelligent system needs to incrementally acquire,
update, accumulate, and exploit knowledge throughout its lifetime. This ability, known as …

Artificial intelligence research has seen enormous progress over the past few decades, but it
predominantly relies on fixed datasets and stationary environments. Continual learning is an …

For future learning systems, incremental learning is desirable because it allows for: efficient
resource usage by eliminating the need to retrain from scratch at the arrival of new data; …

Online continual learning for image classification studies the problem of learning to classify
images from an online stream of data and tasks, where tasks may include new classes …

We discuss a general formulation for the Continual Learning (CL) problem for classification—
a learning task where a stream provides samples to a learner and the goal of the learner …

Artificial neural networks thrive in solving the classification problem for a particular rigid task,
acquiring knowledge through generalized learning behaviour from a distinct training phase …

Modern machine learning suffers from catastrophic forgetting when learning new classes
incrementally. The performance dramatically degrades due to the missing data of old …

We introduce DeepInversion, a new method for synthesizing images from the image
distribution used to train a deep neural network. We" invert" a trained network (teacher) to …

Conventionally, deep neural networks are trained offline, relying on a large dataset
prepared in advance. This paradigm is often challenged in real-world applications, eg online …

Although deep learning approaches have stood out in recent years due to their state-of-the-
art results, they continue to suffer from catastrophic forgetting, a dramatic decrease in overall …