Abstract

Artificial autonomous agents and robots interacting in complex environments are required to continually acquire and fine-tune knowledge over sustained periods of time. The ability to learn from continuous streams of information is referred to as lifelong learning and represents a long-standing challenge for neural network models due to catastrophic forgetting in which novel sensory experience interferes with existing representations and leads to abrupt decreases in the performance on previously acquired knowledge. Computational models of lifelong learning typically alleviate catastrophic forgetting in experimental scenarios with given datasets of static images and limited complexity, thereby differing significantly from the conditions artificial agents are exposed to. In more natural settings, sequential information may become progressively available over time and access to previous experience may be restricted. Therefore, specialized neural network mechanisms are required that adapt to novel sequential experience while preventing disruptive interference with existing representations. In this paper, we propose a dual-memory self-organizing architecture for lifelong learning scenarios. The architecture comprises two growing recurrent networks with the complementary tasks of learning object instances (episodic memory) and categories (semantic memory). Both growing networks can expand in response to novel sensory experience: the episodic memory learns fine-grained spatiotemporal representations of object instances in an unsupervised fashion while the semantic memory uses task-relevant signals to regulate structural plasticity levels and develop more compact representations from episodic experience. For the consolidation of knowledge in the absence of external sensory input, the episodic memory periodically replays trajectories of neural reactivations. We evaluate the proposed model on the CORe50 benchmark dataset for continuous object recognition, showing that we significantly outperform current methods of lifelong learning in three different incremental learning scenarios.

Highlights

  • Artificial autonomous agents and robots interacting in dynamic environments are required to continually acquire and fine-tune their knowledge over time (Thrun and Mitchell, 1995; Parisi et al, 2018a)

  • We reproduce three experimental conditions defined by the CORe50 benchmark showing that our model significantly outperforms state-of-the-art lifelong learning approaches

  • These results suggest that memory replay is an important feature for the reactivation of previously learned neural representations at the moment of learning from novel sensory experience with the goal to prevent that classes that have been encountered at early stages be forgotten over time

Read more

Summary

Introduction

Artificial autonomous agents and robots interacting in dynamic environments are required to continually acquire and fine-tune their knowledge over time (Thrun and Mitchell, 1995; Parisi et al, 2018a). Effective models of lifelong learning are crucial in real-world conditions where an autonomous agent cannot be provided with all the necessary prior knowledge to interact with the environment and the direct access to previous experience is restricted (Thrun and Mitchell, 1995). Lifelong machine learning represents a long-standing challenge due to catastrophic forgetting or interference, i.e., training a model with a new task leads to an abrupt decrease in the performance on previously learned tasks (McCloskey and Cohen, 1989). To overcome catastrophic forgetting, computational models must adapt their existing representations on the basis of novel sensory experience while preventing disruptive interference with previously learned representations. The extent to which a system must be flexible for learning novel knowledge and stable for preventing the disruption of consolidated knowledge is known as the stabilityplasticity dilemma, which has been extensively studied for both computational and biological systems (e.g., Grossberg, 1980, 2007; Mermillod et al, 2013; Ditzler et al, 2015)

Methods
Results
Discussion
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call