Prediction Error-Driven Memory Consolidation for Continual Learning: On the Case of Adaptive Greenhouse Models

Guido Schillaci (1) , Uwe Schmidt (2) , Luis Miranda (2)

1 SSSUP - Scuola Universitaria Superiore Sant'Anna [Pisa]
2 HU Berlin - Humboldt-Universität zu Berlin = Humboldt University of Berlin = Université Humboldt de Berlin

This work presents an adaptive architecture that performs online learning and faces catastrophic forgetting issues by means of an episodic memory system and of prediction-error driven memory consolidation. In line with evidence from brain sciences, memories are retained depending on their congruence with the prior knowledge stored in the system. In this work, congruence is estimated in terms of prediction error resulting from a deep neural model. The proposed AI system is transferred onto an innovative application in the horticulture industry: the learning and transfer of greenhouse models. This work presents models trained on data recorded from research facilities and transferred to a production greenhouse.

Mots clés: Memory consolidation; Greenhouse; Adaptive models; Deep neural networks; Episodic memory

Domaines: Intelligence artificielle [cs.AI]; Informatique

Guido Schillaci, Uwe Schmidt, Luis Miranda. Prediction Error-Driven Memory Consolidation for Continual Learning: On the Case of Adaptive Greenhouse Models. KI - Künstliche Intelligenz, 2021, Developmental Robotics, 35 (1), pp.71 - 80. ⟨10.1007/s13218-020-00700-8⟩. ⟨hal-03344240⟩