Learning system for a data processing apparatus

• US 5,410,636 A
• Filed: 08/27/1993
• Issued: 04/25/1995
• Est. Priority Date: 02/20/1989
• Status: Expired due to Fees

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Masuoka, Ryusuke, Watanabe, Nobuo, Kawamura, Akira, Kimoto, Takashi, Asakawa, Kazuo, Tanahashi, Jun'ichi
• Primary Examiner(s): MacDonald, Allen R.
• Application Number:
  US08/112,377
• Time in Patent Office: 606 Days
• Field of Search: 395/22, 395/23, 395/24
• US Class Current: 706/25
• CPC Class Codes: G06N 3/08 Learning methods

Abstract

A learning system is used in a data processing apparatus for learning an input pattern by obtaining an internal-state value necessary for realizing a desired data conversion by performing a pattern conversion defined by the internal-state value and calculating an output pattern corresponding to the input pattern. The learning system comprises a pattern presenting unit for presenting an input pattern group of the subject to be learned for pattern conversion, dividing the input pattern group of the subject to be learned into at least two sets, selecting one of the divided sets, presenting the input pattern group of the selected set to a pattern conversion unit and presenting an input pattern group belonging to all the sets presented up to the current point when the internal-state value to be converged is obtained in accordance with the presentation of the selected set, and an error value calculating unit for calculating an error value representing a magnitude of a non-consistency between an output pattern group outputted in accordance with the presentation and a teacher pattern group representing a pattern to be obtained by the said output pattern group.

Claims

1. A learning method for use in a data processing apparatus for learning an input pattern by obtaining an internal-state value necessary for realizing a desired data conversion by performing a pattern conversion defined by the internal-state value and calculating an output pattern corresponding to the input pattern, said method comprising the steps of:

(a) presenting an input pattern group of a subject to be learned for pattern conversion by a pattern presenting circuit to the data processing apparatus;

(b) dividing the input pattern group of the subject to be learned into at least two sets;

(c) selecting one of the divided sets;

(d) presenting, by the pattern presenting circuit, the input pattern group of the selected set for pattern conversion and sequentially presenting an input pattern group belonging to all the sets presented up to the current point when internal-state value to be converged is obtained in accordance with the presentation;

(e) calculating an error value, by an error calculating circuit, representing a magnitude of an inconsistency between an output pattern group outputted from a pattern converting neural network, including a multiplier, an accumulator and a threshold value processor, in accordance with the input pattern group and a teacher pattern group representing a pattern to be obtained by the output pattern group; and

(f) updating an internal-state value, by an internal-state value updating circuit, in accordance with an updated quantity of the internal-state value calculated based on the error value from said error calculating circuit and obtaining an internal-state value having an error value within a specified allowance.