A neural net compiler system for hierarchical organization

  • Authors:
  • Rajeev Kumar

  • Affiliations:
  • Indian Institute of Technology, Kharagpur, India

  • Venue:
  • ACM SIGPLAN Notices
  • Year:
  • 2001

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present a language framework for handling arbitrarily complex neural computations. The software architecture - which we call an Artificial Neural Network Compiler for Hierarchical ORganization (ANCHOR) - facilitates network hierarchy and simpler sub-mappings. We define a Net Definition Language (NDL) which is implemented in object-oriented programming paradigm; a trained network is decompiled back into NDL. ANCHOR is configured around the concept of a Superneuron which is a generalized view of a neuron-processing element and designed using reuse of object-model. The indistinguishability between a superneuron and a neuron is employed in hierarchical nesting of superneurons, up to (theoretically) infinite depth within other superneurons as well as linear or tree-structured cascading. Hierarchical decomposition of simple boolean functions has been demonstrated as proof-of-concept.