The ability to learn tasks in a sequential fashion is crucial to the development of artificial intelligence. Until now neural networks have not been capable of this and it has been widely thought that catastrophic forgetting is an inevitable feature of connectionist models. We show that it is possible to overcome this limitation and train networks that can maintain expertise on tasks that they have not experienced for a long time. Our approach remembers old tasks by selectively slowing down learning on the weights important for those tasks. We demonstrate our approach is scalable and effective by solving a set of classification tasks based on a hand-written digit dataset and by learning several Atari 2600 games sequentially.
Download full-text PDF |
Source |
|---|---|
| http://www.ncbi.nlm.nih.gov/pmc/articles/PMC5380101 | PMC |
| http://dx.doi.org/10.1073/pnas.1611835114 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
Continual deep reinforcement learning with task-agnostic policy distillation.
Sci Rep
December 2024
Department of Informatics, University of Hamburg, Hamburg, Germany.
Central to the development of universal learning systems is the ability to solve multiple tasks without retraining from scratch when new data arrives. This is crucial because each task requires significant training time. Addressing the problem of continual learning necessitates various methods due to the complexity of the problem space.
View Article and Find Full Text PDFEnhancing consistency and mitigating bias: A data replay approach for incremental learning.
Neural Netw
December 2024
Department of Automation, Tsinghua University, Beijing 100084, China. Electronic address:
Deep learning systems are prone to catastrophic forgetting when learning from a sequence of tasks, as old data from previous tasks is unavailable when learning a new task. To address this, some methods propose replaying data from previous tasks during new task learning, typically using extra memory to store replay data. However, it is not expected in practice due to memory constraints and data privacy issues.
View Article and Find Full Text PDFAn optimized BERT fine-tuned model using an artificial bee colony algorithm for automatic essay score prediction.
PeerJ Comput Sci
September 2024
Faculty of Information Science and Technology, Universiti Kebangsaan Malaysia, Bangi, Selangor, Malaysia.
Background: The Automatic Essay Score (AES) prediction system is essential in education applications. The AES system uses various textural and grammatical features to investigate the exact score value for AES. The derived features are processed by various linear regressions and classifiers that require the learning pattern to improve the overall score.
View Article and Find Full Text PDFLabel-Guided relation prototype generation for Continual Relation Extraction.
PeerJ Comput Sci
October 2024
Faculty of Electrical Engineering and Computer Science, University of Maribor, Maribor, Slovenia.
Continual relation extraction (CRE) aims to extract relations towards the continuous and iterative arrival of new data. To address the problem of catastrophic forgetting, some existing research endeavors have focused on exploring memory replay methods by storing typical historical learned instances or embedding all observed relations as prototypes by averaging the hidden representation of samples and replaying them in the subsequent training process. However, this prototype generation method overlooks the rich semantic information within the label namespace and are also constrained by memory size, resulting in inadequate descriptions of relation semantics by relation prototypes.
View Article and Find Full Text PDFMetaplasticity-Enabled Graphene Quantum Dot Devices for Mitigating Catastrophic Forgetting in Artificial Neural Networks.
Adv Mater
December 2024
School of Integrated Circuits, Zhejiang University, Hangzhou, Zhejiang, 311200, China.
The limitations of deep neural networks in continuous learning stem from oversimplifying the complexities of biological neural circuits, often neglecting the dynamic balance between memory stability and learning plasticity. In this study, artificial synaptic devices enhanced with graphene quantum dots (GQDs) that exhibit metaplasticity is introduced, a higher-order form of synaptic plasticity that facilitates the dynamic regulation of memory and learning processes similar to those observed in biological systems. The GQDs-assisted devices utilize interface-mediated modifications in asymmetric conductive pathways, replicating classical synaptic plasticity mechanisms.
View Article and Find Full Text PDFWant AI Summaries of new PubMed Abstracts delivered to your In-box?
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!