Abstract
Recently, many researches on associative memories have been made and a lot of neural network models have been proposed. Bidirectional Associative Memory (BAM) is one of them. The BAM uses Hebbian learning. However, unless the training vectors are orthogonal, Hebbian learning does not guarantee the recall of all training pairs. Namely, the BAM which is trained by Hebbian learning suffers from low memory capacity. To improve the storage capacity of the BAM, Pseudo-Relaxation Learning Algorithm for BAM (PRLAB) has been proposed. However, PRLAB needs long learning epochs because of random initial weights. In this paper, we propose Quick Learning for BAM which greatly reduces learning epochs and guarantees the recall of all training pairs. In the proposed algorithm, the BAM is trained by Hebbian learning in the first stage and then trained by PRLAB. Owing to the use of Hebbian learning in the first stage, the weights are much closer to the solution space than the initial weights chosen randomly. As a result, the proposed algorithm can reduce the learning epochs. The features of the proposed algorithm are: 1) It requires much less learning epochs. 2) It guarantees the recall of all training pairs. 3) It is robust for noisy inputs. 4) The memory capacity is much larger than conventional BAM. In addition, we made clear several important characteristics of the conventional and the proposed algorithms such as noise reduction characteristics, storage capacity and the finding of an index which relates to the noise reduction.
Original language | English |
---|---|
Pages (from-to) | 385-392 |
Number of pages | 8 |
Journal | IEICE Transactions on Information and Systems |
Volume | E77-D |
Issue number | 4 |
Publication status | Published - 1994 Apr |
Fingerprint
ASJC Scopus subject areas
- Computer Graphics and Computer-Aided Design
- Information Systems
- Software
Cite this
Quick learning for bidirectional associative memory. / Hattori, Motonobu; Hagiwara, Masafumi; Nakagawa, Masao.
In: IEICE Transactions on Information and Systems, Vol. E77-D, No. 4, 04.1994, p. 385-392.Research output: Contribution to journal › Article
}
TY - JOUR
T1 - Quick learning for bidirectional associative memory
AU - Hattori, Motonobu
AU - Hagiwara, Masafumi
AU - Nakagawa, Masao
PY - 1994/4
Y1 - 1994/4
N2 - Recently, many researches on associative memories have been made and a lot of neural network models have been proposed. Bidirectional Associative Memory (BAM) is one of them. The BAM uses Hebbian learning. However, unless the training vectors are orthogonal, Hebbian learning does not guarantee the recall of all training pairs. Namely, the BAM which is trained by Hebbian learning suffers from low memory capacity. To improve the storage capacity of the BAM, Pseudo-Relaxation Learning Algorithm for BAM (PRLAB) has been proposed. However, PRLAB needs long learning epochs because of random initial weights. In this paper, we propose Quick Learning for BAM which greatly reduces learning epochs and guarantees the recall of all training pairs. In the proposed algorithm, the BAM is trained by Hebbian learning in the first stage and then trained by PRLAB. Owing to the use of Hebbian learning in the first stage, the weights are much closer to the solution space than the initial weights chosen randomly. As a result, the proposed algorithm can reduce the learning epochs. The features of the proposed algorithm are: 1) It requires much less learning epochs. 2) It guarantees the recall of all training pairs. 3) It is robust for noisy inputs. 4) The memory capacity is much larger than conventional BAM. In addition, we made clear several important characteristics of the conventional and the proposed algorithms such as noise reduction characteristics, storage capacity and the finding of an index which relates to the noise reduction.
AB - Recently, many researches on associative memories have been made and a lot of neural network models have been proposed. Bidirectional Associative Memory (BAM) is one of them. The BAM uses Hebbian learning. However, unless the training vectors are orthogonal, Hebbian learning does not guarantee the recall of all training pairs. Namely, the BAM which is trained by Hebbian learning suffers from low memory capacity. To improve the storage capacity of the BAM, Pseudo-Relaxation Learning Algorithm for BAM (PRLAB) has been proposed. However, PRLAB needs long learning epochs because of random initial weights. In this paper, we propose Quick Learning for BAM which greatly reduces learning epochs and guarantees the recall of all training pairs. In the proposed algorithm, the BAM is trained by Hebbian learning in the first stage and then trained by PRLAB. Owing to the use of Hebbian learning in the first stage, the weights are much closer to the solution space than the initial weights chosen randomly. As a result, the proposed algorithm can reduce the learning epochs. The features of the proposed algorithm are: 1) It requires much less learning epochs. 2) It guarantees the recall of all training pairs. 3) It is robust for noisy inputs. 4) The memory capacity is much larger than conventional BAM. In addition, we made clear several important characteristics of the conventional and the proposed algorithms such as noise reduction characteristics, storage capacity and the finding of an index which relates to the noise reduction.
UR - http://www.scopus.com/inward/record.url?scp=0028410032&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0028410032&partnerID=8YFLogxK
M3 - Article
AN - SCOPUS:0028410032
VL - E77-D
SP - 385
EP - 392
JO - IEICE Transactions on Information and Systems
JF - IEICE Transactions on Information and Systems
SN - 0916-8532
IS - 4
ER -