Download PDFOpen PDF in browser

Artificial Superintelligence : An AI That Makes Better AI’s Recursively.

EasyChair Preprint no. 4077

11 pagesDate: August 25, 2020


Artificial Super Intelligence or ASI that is more potent and refined than human’s intelligence. ASI is based on the ideas that machines can imitate the human mind, their way of working to the extent that they can even supersede them. As a first step, ASI aims to improve the intelligent abilities of the machines and to achieve this, the ASI will have to make an AI which makes better AI’s recursively for achieving high-level intelligence. In this paper, we discuss classic R-CNN model with different SVM architectures/algorithms for object detection implementation to get varying outcomes which in turn results in better AI’s with varying degree of accuracy recursively. However, for simplicity we have taken CNN-SVM combination where an AI algorithm( CNN ) passes intelligence to fast machine learning( SVM ) algorithm recursively for solving multiclass problems from large data sets that implements object detection for designing a better AI machine. The test results are encouraging with high accuracy and the model is therefore shown that a lesser AI is making a better AI when combined with intelligent vector algorithm recursively resulting in very high level of intelligence.

Keyphrases: artificial superintelligence, CNN-SVM Architecture, Recursive AI, Recursive Self-Improvement

BibTeX entry
BibTeX does not have the right entry for preprints. This is a hack for producing the correct reference:
  author = {Poondru Prithvinath Reddy},
  title = {Artificial Superintelligence : An AI That Makes Better AI’s Recursively.},
  howpublished = {EasyChair Preprint no. 4077},

  year = {EasyChair, 2020}}
Download PDFOpen PDF in browser