KR970051806A - 스레숄드 망의 합성방법 - Google Patents

스레숄드 망의 합성방법 Download PDF

Info

Publication number
KR970051806A
KR970051806A KR1019950053633A KR19950053633A KR970051806A KR 970051806 A KR970051806 A KR 970051806A KR 1019950053633 A KR1019950053633 A KR 1019950053633A KR 19950053633 A KR19950053633 A KR 19950053633A KR 970051806 A KR970051806 A KR 970051806A
Authority
KR
South Korea
Prior art keywords
threshold
network
layer
synthesizing
convergence
Prior art date
Application number
KR1019950053633A
Other languages
English (en)
Inventor
김정환
박승권
한영남
Original Assignee
양승택
한국전자통신연구원
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 양승택, 한국전자통신연구원 filed Critical 양승택
Priority to KR1019950053633A priority Critical patent/KR970051806A/ko
Priority to US08/708,544 priority patent/US5980079A/en
Publication of KR970051806A publication Critical patent/KR970051806A/ko

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/082—Learning methods modifying the architecture, e.g. adding, deleting or silencing nodes or connections
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Image Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

본 발명은 어떠한 절체함수 하에서도 자동적으로 숨겨진 층의 요구되는 스레숄드(thershold) 요소의 수를 결정하며 확실하게 수렴되는 3층의 스레숄드 망을 위한 학습 알고리듬을 통하여 디지털 VLSI 기술로 제안된 3층의 스레숄드 망의 합성방법에 관한 것으로서, 그 특징은 스래숄드 망의 합성방법에 있어서, 망 입력과 출력을 갖는 3계층의 스레숄드 망을 제공하고, 결정된 숨겨진 계층 내에서 요구되는 스레숄드 요소의 수를 가지는 수렴이 보장된 각 스레숄드 요소의 가중치와 스레숄드의 집합을 보장하는 기하학적 확장 학습알고리듬을 포함하여, 어떠한 함수에 대해서도 수렴이 보장되어지며, 3계층으로 된 데에 있으므로, 본 발명은 어떠한 절체함수하에서도 자동적으로 숨겨진 층의 요구되는 스레숄드 요소의 수를 결정하며 확실하게 수렴되는 3층의 스레숄드 망을 학습알고리듬을 통하여 VLSI 기술로 제안된 3층의 스레숄드 망을 구축할 수 있다는 데에 그 효과가 있다.

Description

스레숄드 망의 합성방법
본 내용은 요부공개 건이므로 전문내용을 수록하지 않았음
제1도는 본 발명에 따른 3층의 스레숄드 망을 도시한 도면.

Claims (1)

  1. 스레숄드 망의 합성방법에 있어서, 망 입력과 출력을 갖는 3계층의 스레숄드 망을 제공하고; 결정된 숨겨진 계층 내에서 요구되는 스레숄드 요소의 수를 가지는 수렴이 보장된 각 스레숄드 요소의 가중치와 스레숄드의 집합을 보장하는 기하학적 확장 학습 알고리듬을 포함하여; 어떠한 함수에 대해서도 수렴이 보장되어지며; 3계층으로 된 것을 특징으로 하는 스레숄드 망의 합성방법.
    ※ 참고사항 : 최초출원 내용에 의하여 공개하는 것임.
KR1019950053633A 1995-12-21 1995-12-21 스레숄드 망의 합성방법 KR970051806A (ko)

Priority Applications (2)

Application Number Priority Date Filing Date Title
KR1019950053633A KR970051806A (ko) 1995-12-21 1995-12-21 스레숄드 망의 합성방법
US08/708,544 US5980079A (en) 1995-12-21 1996-09-05 Synthesis of threshold networks

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1019950053633A KR970051806A (ko) 1995-12-21 1995-12-21 스레숄드 망의 합성방법

Publications (1)

Publication Number Publication Date
KR970051806A true KR970051806A (ko) 1997-07-29

Family

ID=19442520

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1019950053633A KR970051806A (ko) 1995-12-21 1995-12-21 스레숄드 망의 합성방법

Country Status (2)

Country Link
US (1) US5980079A (ko)
KR (1) KR970051806A (ko)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9262726B2 (en) * 2013-01-17 2016-02-16 Applied Materials, Inc. Using radial basis function networks and hyper-cubes for excursion classification in semi-conductor processing equipment

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE69031842T2 (de) * 1989-02-20 1998-04-16 Fujitsu Ltd Lernsystem und Lernverfahren für eine Datenverarbeitungsvorrichtung
JPH0438587A (ja) * 1990-06-04 1992-02-07 Nec Corp 入力領域適応型ニューラルネットワーク文字認識装置
JPH0683792A (ja) * 1991-06-12 1994-03-25 Hitachi Ltd ニューラルネットワークの学習装置およびニューラルネットワークの学習パターン呈示方法
US5353207A (en) * 1992-06-10 1994-10-04 Pavilion Technologies, Inc. Residual activation neural network

Also Published As

Publication number Publication date
US5980079A (en) 1999-11-09

Similar Documents

Publication Publication Date Title
Holmes et al. Non‐Expansive Actions of Topological Semigroups and Fixed Points
GB0007318D0 (en) A system for generating musical sounds
DK0722644T3 (da) Fremgangsmåde til udvælgelse af forbindelser i net
GB9503760D0 (en) Neural networks
DE69114640D1 (de) Biaxial orientierte Verbundfolie.
DE69636165D1 (de) Mehrschichtiges minimallogiknetzwerk
DE69607153T2 (de) Kriechbeständiger formkörper aus dichtem, expandiertem ptfe
ID16084A (id) Unsur-unsur ptfe yang diperpanjang dan bagian-bagian yang dibuatnya.
KR970051806A (ko) 스레숄드 망의 합성방법
MY115094A (en) Aluminum mesh with interlaced hollow and solid ribs
Shastri Types and quantifiers in SHRUTI–a connectionist model of rapid reasoning and relational processing
KR970051805A (ko) 새로운 웨이퍼 홀더(wafer holder)를 구비하는 반도체 제조장치
Kiefer Thematic roles and compounds
Thomas et al. Infinite generation of non-cocompact lattices on right-angled buildings
Shahgholian A characterization of the sphere in terms of single-layer potentials
Rothschild Further notes on the allocation of effort
Mares CE is not a conservative extension of E
Bartoszyński et al. All meager filters may be null
Popa Irresolute multifunctions
KR920013171A (ko) 신경회로망을 위한 프로세서 어레이 구조
Cline The Tempest by Shakespeare's Globe
VERSTRAETE ASSIGNMENT OF FUNCTIONAL RESPONSIBILITY IN PERCEPTRONS (LEARNING)
Nguyen et al. ON FUZZY INFERENCE BASED ON α-LEVEL SETS
Koeppel Designing Gotham: West Point Engineers and the Rise of Modern New York, 1817–1898 by Jon Scott Logel
Cubiotti A counter-example on a quasi-variational inequality without lower semicontinuity assumption

Legal Events

Date Code Title Description
A201 Request for examination
E902 Notification of reason for refusal
E601 Decision to refuse application