Abstract
It is widely known that reinforcement learning is a more general problem than supervised learning. In fact, supervised learning can be seen as a class of reinforcement learning problems. However, only a couple of papers tested reinforcement learning algorithms in supervised learning problems. Here we propose a new and simpler way to abstract supervised learning for any reinforcement learning algorithm. Moreover, a new algorithm called Novelty-Organizing Classifiers is developed based on a Novelty Map population that focuses more on the novelty of the inputs than their frequency. A comparison of the proposed method with Self-Organizing Classifiers and BioHel on some datasets is presented. Even though BioHel is specialized in solving supervised learning problems, the results showed only a trade-off between the algorithms. Lastly, results on a maze problem validate the flexibility of the proposed algorithm beyond supervised learning problems. Thus, Novelty-Organizing Classifiers is capable of solving many supervised learning problems as well as a maze problem without changing any parameter at all. Considering the fact that no adaptation of parameters was executed, the proposed algorithm's basis seems interestingly flexible.
| Original language | English |
|---|---|
| Title of host publication | GECCO 2014 - Companion Publication of the 2014 Genetic and Evolutionary Computation Conference |
| Publisher | Association for Computing Machinery |
| Pages | 81-82 |
| Number of pages | 2 |
| ISBN (Print) | 9781450328814 |
| DOIs | |
| Publication status | Published - 2014 |
| Event | 16th Genetic and Evolutionary Computation Conference, GECCO 2014 - Vancouver, BC, Canada Duration: Jul 12 2014 → Jul 16 2014 |
Other
| Other | 16th Genetic and Evolutionary Computation Conference, GECCO 2014 |
|---|---|
| Country | Canada |
| City | Vancouver, BC |
| Period | 7/12/14 → 7/16/14 |
Fingerprint
All Science Journal Classification (ASJC) codes
- Computational Theory and Mathematics
- Applied Mathematics
Cite this
Novelty-Organizing Classifiers applied to classification and reinforcement learning : Towards flexible algorithms. / Vargas, Danilo Vasconcellos; Takano, Hirotaka; Murata, Junichi.
GECCO 2014 - Companion Publication of the 2014 Genetic and Evolutionary Computation Conference. Association for Computing Machinery, 2014. p. 81-82.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Novelty-Organizing Classifiers applied to classification and reinforcement learning
T2 - Towards flexible algorithms
AU - Vargas, Danilo Vasconcellos
AU - Takano, Hirotaka
AU - Murata, Junichi
PY - 2014
Y1 - 2014
N2 - It is widely known that reinforcement learning is a more general problem than supervised learning. In fact, supervised learning can be seen as a class of reinforcement learning problems. However, only a couple of papers tested reinforcement learning algorithms in supervised learning problems. Here we propose a new and simpler way to abstract supervised learning for any reinforcement learning algorithm. Moreover, a new algorithm called Novelty-Organizing Classifiers is developed based on a Novelty Map population that focuses more on the novelty of the inputs than their frequency. A comparison of the proposed method with Self-Organizing Classifiers and BioHel on some datasets is presented. Even though BioHel is specialized in solving supervised learning problems, the results showed only a trade-off between the algorithms. Lastly, results on a maze problem validate the flexibility of the proposed algorithm beyond supervised learning problems. Thus, Novelty-Organizing Classifiers is capable of solving many supervised learning problems as well as a maze problem without changing any parameter at all. Considering the fact that no adaptation of parameters was executed, the proposed algorithm's basis seems interestingly flexible.
AB - It is widely known that reinforcement learning is a more general problem than supervised learning. In fact, supervised learning can be seen as a class of reinforcement learning problems. However, only a couple of papers tested reinforcement learning algorithms in supervised learning problems. Here we propose a new and simpler way to abstract supervised learning for any reinforcement learning algorithm. Moreover, a new algorithm called Novelty-Organizing Classifiers is developed based on a Novelty Map population that focuses more on the novelty of the inputs than their frequency. A comparison of the proposed method with Self-Organizing Classifiers and BioHel on some datasets is presented. Even though BioHel is specialized in solving supervised learning problems, the results showed only a trade-off between the algorithms. Lastly, results on a maze problem validate the flexibility of the proposed algorithm beyond supervised learning problems. Thus, Novelty-Organizing Classifiers is capable of solving many supervised learning problems as well as a maze problem without changing any parameter at all. Considering the fact that no adaptation of parameters was executed, the proposed algorithm's basis seems interestingly flexible.
UR - http://www.scopus.com/inward/record.url?scp=84905667096&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84905667096&partnerID=8YFLogxK
U2 - 10.1145/2598394.2598429
DO - 10.1145/2598394.2598429
M3 - Conference contribution
AN - SCOPUS:84905667096
SN - 9781450328814
SP - 81
EP - 82
BT - GECCO 2014 - Companion Publication of the 2014 Genetic and Evolutionary Computation Conference
PB - Association for Computing Machinery
ER -