Abstract | ||
---|---|---|
Some real-world domains are best characterized as a single task, but for others this perspective is limiting. Instead, some tasks continually grow in complexity, in tandem with the agentu0027s competence. In continual learning, also referred to as lifelong learning, there are no explicit task boundaries or curricula. As learning agents have become more powerful, continual learning remains one of the frontiers that has resisted quick progress. To test continual learning capabilities we consider a challenging 3D domain with an implicit sequence of tasks and sparse rewards. We propose a novel agent architecture called Unicorn, which demonstrates strong continual learning and outperforms several baseline agents on the proposed domain. The agent achieves this by jointly representing and learning multiple policies efficiently, using a parallel off-policy learning setup. |
Year | Venue | Field |
---|---|---|
2018 | arXiv: Learning | Unicorn,Agent architecture,Human–computer interaction,Curriculum,Artificial intelligence,Lifelong learning,Machine learning,Mathematics,Limiting |
DocType | Volume | Citations |
Journal | abs/1802.08294 | 2 |
PageRank | References | Authors |
0.37 | 29 | 10 |
Name | Order | Citations | PageRank |
---|---|---|---|
Daniel J. Mankowitz | 1 | 29 | 8.05 |
Augustin Zídek | 2 | 10 | 1.16 |
André Barreto | 3 | 2 | 2.40 |
Dan Horgan | 4 | 105 | 4.38 |
Matteo Hessel | 5 | 133 | 10.65 |
john quan | 6 | 339 | 13.28 |
junhyuk oh | 7 | 235 | 13.79 |
hado van hasselt | 8 | 432 | 31.39 |
David Silver | 9 | 8252 | 363.86 |
Tom Schaul | 10 | 916 | 79.40 |