A Theoretical Analysis of On-Line Learning Using Correlated Examples

Seki, Chihiro; Sakurai, Shingo; Matsuno, Masafumi; Miyoshi, Seiji

doi:10.1093/ietfec/e91-a.9.2663

Cited by 3 publications

(3 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, the block orthogonal projection learning was proposed as a learning rule whose learning speed is not affected by the correlation [3]. In the case of nonlinear perceptrons, the different behaviors described in this section are exhibited by different learning rules.…”

Section: Resultsmentioning

confidence: 99%

“…However, such correlation is difficult to treat analytically. Therefore, in this paper we treat the correlation considered in [3].…”

Section: Modelmentioning

confidence: 99%

“…In practical applications, inputs are considered to be mutually correlated in many cases. The treatment of correlated inputs by a model composed of linear perceptrons has already been analyzed [3]. Considering the applications of pattern recognition, it is important to theoretically investigate the generalization performance of a nonlinear learning machine using correlated inputs.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Statistical Mechanics of On-Line Learning Using Correlated Examples

Nakao

Narukawa

Miyoshi

2011

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYWe consider a model composed of nonlinear perceptrons and analytically investigate its generalization performance using correlated examples in the framework of on-line learning by a statistical mechanical method. In Hebbian and AdaTron learning, the larger the number of examples used in an update, the slower the learning. In contrast, Perceptron learning does not exhibit such behaviors, and the learning becomes fast in some time region.

show abstract

Section: Resultsmentioning

confidence: 99%

“…However, such correlation is difficult to treat analytically. Therefore, in this paper we treat the correlation considered in [3].…”

Section: Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation