Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

A parallel neural network simulator on the connection machine CM-5

M Reczko1, A Hatzigeorgiou, N Mache

  • 1Molecular Biophysics Department, German Cancer Research Centre, Heidelberg.

Computer Applications in the Biosciences : CABIOS
|June 1, 1995
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Knight Scholars Program: A Tiered Three-Year Mentored Training Program for Urban and Rural High School High School Students Increases Interest and Self-Efficacy in Interprofessional Cancer Research.

Journal of STEM outreach·2023
Same author

Interprofessional Near-Peer Mentoring Teams Enhance Cancer Research Training: Sustainable Approaches for Biomedical Workforce Development of Historically Underrepresented Students.

Journal of STEM outreach·2023
Same author

Neuroprotection requires the functions of the RNA-binding protein HuR.

Cell death and differentiation·2014
Same author

Investigating recurrent neural networks for OCT A-scan based tissue analysis.

Methods of information in medicine·2014
Same author

Genomics of ADME gene expression: mapping expression quantitative trait loci relevant for absorption, distribution, metabolism and excretion of drugs in human liver.

The pharmacogenomics journal·2011
Same author

Activation of the mitogen-activated protein kinase (MAPK) signalling pathway in the liver of mice is related to plasma glucose levels after acute exercise.

Diabetologia·2010
Same journal

DCA: an efficient implementation of the divide-and-conquer approach to simultaneous multiple sequence alignment.

Computer applications in the biosciences : CABIOS·1998
Same journal

Two applications to facilitate the viewing of database search result files on the Macintosh.

Computer applications in the biosciences : CABIOS·1998
Same journal

BioWish: a molecular biology command extension to Tcl/Tk.

Computer applications in the biosciences : CABIOS·1998
Same journal

The Sequence Alerting Server--a new WEB server.

Computer applications in the biosciences : CABIOS·1998
Same journal

A software tool for the analysis of mass spectrometric disulfide mapping experiments.

Computer applications in the biosciences : CABIOS·1998
Same journal

SAMBA: hardware accelerator for biological sequence comparison.

Computer applications in the biosciences : CABIOS·1998
See all related articles

This study presents a scalable parallel artificial neural network implementation on the CM-5, achieving high performance for DNA sequence analysis. The efficient design enables rapid training of large neural networks for molecular biology applications.

Area of Science:

  • Computer Science
  • Computational Biology
  • Bioinformatics

Background:

  • Artificial neural networks (ANNs) are powerful tools for complex pattern recognition.
  • Training large ANNs with extensive datasets is computationally intensive.
  • Efficient parallel implementations are crucial for advancing applications in fields like molecular biology.

Purpose of the Study:

  • To develop and evaluate a parallel implementation of artificial neural networks on the Connection Machine CM-5.
  • To assess the scalability and performance of this implementation for training large neural networks.
  • To apply the implementation to the prediction of coding regions in DNA sequences.

Main Methods:

  • Developed a parallel ANN simulator utilizing training pattern parallelism on the CM-5.

Related Experiment Videos

  • Leveraged CM-5's parallel I/O and control network reduction operations for scalability.
  • Compared performance against other Single Instruction, Multiple Data (SIMD) and Multiple Instruction, Multiple Data (MIMD) architectures.
  • Tested with various batch training algorithms including Backpropagation (BP), Quickprop, and Rprop.

    • Main Results:

    • Achieved a maximum speed of 149.25 Million Connections Per Second (MCUPS) on a 512-processor CM-5 system.
    • Demonstrated high scalability due to efficient use of CM-5's parallel features.
    • The implementation supports unrestricted network topologies and multiple training algorithms.

    Conclusions:

    • The parallel ANN implementation on the CM-5 offers significant performance gains for large-scale training.
    • This approach is highly effective for computationally demanding tasks in molecular biology, such as DNA sequence analysis.
    • The method provides a flexible and scalable solution for training diverse neural network architectures.