In a standard view of the reinforcement learning problem, an agent's goal is to efficiently
identify a policy that maximizes long-term reward. However, this perspective is based on a …

Underpinning the past decades of work on the design, initialization, and optimization of
neural networks is a seemingly innocuous assumption: that the network is trained on a\textit …

Continual learning is a sub-field of machine learning, which aims to allow machine learning
models to continuously learn on new data, by accumulating knowledge without forgetting …

Modern deep-learning systems are specialized to problem settings in which training occurs
once and then never again, as opposed to continual-learning settings in which training …

Abstract Training Generative Adversarial Networks (GANs) remains a challenging problem.
The discriminator trains the generator by learning the distribution of real/generated data …

This study investigates the loss of generalization ability in neural networks, revisiting warm-
starting experiments from Ash & Adams. Our empirical analysis reveals that common …

Modern reinforcement learning has been conditioned by at least three dogmas. The first is
the environment spotlight, which refers to our tendency to focus on modeling environments …

Many failures in deep continual and reinforcement learning are associated with increasing
magnitudes of the weights, making them hard to change and potentially causing overfitting …

A sequential decision-making agent balances between exploring to gain new knowledge
about an environment and exploiting current knowledge to maximize immediate reward. For …

In continual learning, knowledge must be preserved and re-used between tasks,
maintaining good transfer to future tasks and minimizing forgetting of previously learned …