Continual Learning through Evolvable Neural Turing Machines

Benno Lüders; Mikkel Schläger; Sebastian Risi

Continual Learning through Evolvable Neural Turing Machines

Benno Lüders, Mikkel Schläger, Sebastian Risi

Research output: Contribution to conference - NOT published in proceeding or journal › Paper › Research › peer-review

Abstract

Continual learning, i.e. the ability to sequentially learn tasks without catastrophic
forgetting of previously learned ones, is an important open challenge in machine
learning. In this paper we take a step in this direction by showing that the recently
proposed Evolving Neural Turing Machine (ENTM) approach is able to perform
one-shot learning in a reinforcement learning task without catastrophic forgetting
of previously stored associations.

Original language	English
Publication date	2016
Number of pages	5
Publication status	Published - 2016

Keywords

Continual learning
Catastrophic forgetting
Evolving Neural Turing Machine
One-shot learning
Reinforcement learning

Access to Document

CLDL-2016_paper_9Accepted author manuscript, 526 KB

https://sites.google.com/site/cldlnips2016/submissions

Cite this

@conference{c280bad8f75b4410bdb2f8c102d7ff85,

title = "Continual Learning through Evolvable Neural Turing Machines",

abstract = "Continual learning, i.e. the ability to sequentially learn tasks without catastrophicforgetting of previously learned ones, is an important open challenge in machinelearning. In this paper we take a step in this direction by showing that the recentlyproposed Evolving Neural Turing Machine (ENTM) approach is able to performone-shot learning in a reinforcement learning task without catastrophic forgettingof previously stored associations.",

keywords = "Continual learning, Catastrophic forgetting, Evolving Neural Turing Machine, One-shot learning, Reinforcement learning, Continual learning, Catastrophic forgetting, Evolving Neural Turing Machine, One-shot learning, Reinforcement learning",

author = "Benno L{\"u}ders and Mikkel Schl{\"a}ger and Sebastian Risi",

note = "Best paper award",

year = "2016",

language = "English",

}

TY - CONF

T1 - Continual Learning through Evolvable Neural Turing Machines

AU - Lüders, Benno

AU - Schläger, Mikkel

AU - Risi, Sebastian

N1 - Best paper award

PY - 2016

Y1 - 2016

N2 - Continual learning, i.e. the ability to sequentially learn tasks without catastrophicforgetting of previously learned ones, is an important open challenge in machinelearning. In this paper we take a step in this direction by showing that the recentlyproposed Evolving Neural Turing Machine (ENTM) approach is able to performone-shot learning in a reinforcement learning task without catastrophic forgettingof previously stored associations.

AB - Continual learning, i.e. the ability to sequentially learn tasks without catastrophicforgetting of previously learned ones, is an important open challenge in machinelearning. In this paper we take a step in this direction by showing that the recentlyproposed Evolving Neural Turing Machine (ENTM) approach is able to performone-shot learning in a reinforcement learning task without catastrophic forgettingof previously stored associations.

KW - Continual learning

KW - Catastrophic forgetting

KW - Evolving Neural Turing Machine

KW - One-shot learning

KW - Reinforcement learning

KW - Continual learning

KW - Catastrophic forgetting

KW - Evolving Neural Turing Machine

KW - One-shot learning

KW - Reinforcement learning

M3 - Paper

ER -

Continual Learning through Evolvable Neural Turing Machines

Abstract

Keywords

Access to Document

Fingerprint

Cite this