Accession Number:

AD1183504

Title:

PySigma: Towards Enhanced Grand Unification for the Sigma Cognitive Architecture

Descriptive Note:

[Technical Report, Research Paper]

Corporate Author:

UNIVERSITY OF SOUTHERN CALIFORNIA LOS ANGELES

Personal Author(s):

• Zhou, Jincheng
• Ustun, Volkan

Report Date:

2021-02-06

Pagination or Media Count:

12

Abstract:

The Sigma cognitive architecture is the beginning of an integrated computational model of intelligent behavior aimed at the grand goal of artificial general intelligence AGI. However, whereas it has been proven to be capable of modeling a wide range of intelligent behaviors, the existing implementation of Sigma has suffered from several significant limitations. The most prominent one is the inadequate support for inference and learning on continuous variables. In this article, we propose solutions for this limitation that should together enhance Sigmas level of grand unification that is, its ability to span both traditional cognitive capabilities and key non-cognitive capabilities central to general intelligence, bridging the gap between symbolic, probabilistic, and neural processing. The resulting design changes converge on a more capable version of the architecture called PySigma. We demonstrate such capabilities of PySigma in neural probabilistic processing via deep generative models, specifically variational autoencoders, as a concrete example.

Descriptors:

• neural networks
• bayesian networks
• cognitive science
• computational science
• monte carlo method
• probabilistic models
• signal processing
• artificial intelligence
• information processing
• information systems
• random variables
• computer science
• information science
• machine learning
• probability
• algorithms
• deep learning

Subject Categories:

• Computer Systems

Distribution Statement:

[A, Approved For Public Release]