Autonomous Programming for General Purposes: Theory and Experiments

Weng, Juyang; Zheng, Zejia; Wu, Xiang; Castro-Garcia, Juan L.

doi:10.1109/ijcnn48605.2020.9207149

Cited by 3 publications

(6 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For further detail of learning vision-guided navigation and planning for navigation see [35]. Below, we will see that an auditory consciousness also uses the same DN, but using different "innate" parameters.…”

Section: Methodsmentioning

confidence: 99%

Conscious Intelligence Requires Lifelong Autonomous Programming For General Purposes

Weng

2020

Preprint

Self Cite

View full text Add to dashboard Cite

Universal Turing Machines [29,10,18] are well known in computer science but they are about manual programming for general purposes. Although human children perform conscious learning (learning while being conscious) from infancy [24,23,14,4], it is unknown that Universal Turing Machiness can facilitate not only our understanding of Autonomous Programming For General Purposes (APFGP) by machines, but also enable early-age conscious learning. This work reports a new kind of AI-conscious learning AI from a machine's "baby" time. Instead of arguing what static tasks a conscious machine should be able to do during its "adulthood", this work suggests that APFGP is a computationally clearer and necessary criterion for us to judge whether a machine is capable of conscious learning so that it can autonomously acquire skills along its "career path". The results here report new concepts and experimental studies for early vision, audition, natural language understanding, and emotion, with conscious learning capabilities that are absent from traditional AI systems.

show abstract

Section: Methodsmentioning

confidence: 99%

Conscious Intelligence Requires Lifelong Autonomous Programming For General Purposes

Weng

2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…( 1) and very important, the space Z(t) is the directly teachable space for the learning system, inspired by brains [50][51][52][53][54]. This new formulation is meant to model not only the brain's spatial processing [55] and temporal processing [56], but also Autonomous Programming for General Purposes (APFPG) [57,58]. Based on the APFGP capability, the AI field seems to have a powerful yet generalpurpose framework toward conscious machines [59][60][61][62].…”

Section: Developmental Schoolmentioning

confidence: 99%

“…The competing standard of the ImageNet competitions [71] did not include any of these four conditions. The AIML Contests [72] considered all the Four Learning Conditions in performance evaluation. In the following subsection, we discuss why task-nonspecificity and incremental mode should be considered in any comparisons.…”

Section: The Four Learning Conditionsmentioning

confidence: 99%

See 1 more Smart Citation

Perspective Chapter: Deep Learning Misconduct and How Conscious Learning Avoids It

Weng

2024

Artificial Intelligence

Self Cite

View full text Add to dashboard Cite

“Deep learning” uses Post-Selection—selection of a model after training multiple models using data. The performance data of “Seep Learning” have been deceptively inflated due to two misconducts: 1: cheating in the absence of a test; 2: hiding bad-looking data. Through the same misconducts, a simple method Pure-Guess Nearest Neighbor (PGNN) gives no errors on any validation dataset V, as long as V is in the possession of the authors and both the amount of storage space and the time of training are finite but unbounded. The misconducts are fatal, because “Deep Learning” is not generalizable, by overfitting a sample set V. The charges here are applicable to all learning modes. This chapter proposes new AI metrics, called developmental errors for all networks trained, under four Learning Conditions: (1) a body including sensors and effectors, (2) an incremental learning architecture (due to the “big data” flaw), (3) a training experience, and (4) a limited amount of computational resources. Developmental Networks avoid Deep Learning misconduct because they train a sole system, which automatically discovers context rules on the fly by generating emergent Turing machines that are optimal in the sense of maximum likelihood across a lifetime, conditioned on the four Learning Conditions.

show abstract

“…Acknowledgements. The author likes to thank Zejia Zheng, Xiang Wu and Juan Castro-Garcia for conducting the experiments, vision, audition, and natural language acquisition, respectively, outlined here with more details reported in [31].…”

mentioning

confidence: 99%

Machines Develop Consciousness Through Autonomous Programming for General Purposes (APFGP)

Weng

2021

Communications in Computer and Information Science

Self Cite

View full text Add to dashboard Cite

Consider a question, "Can machines be conscious?" The subject "consciousness" is vague and challenging. Although there has been a rich collection of literature on consciousness, computational modeling of consciousness that is both holistic in scope and detailed in simulatable computation is lacking. Based on recent advances on a new capability-Autonomous Programming For General Purposes (APFGP)-this work presents APFGP as a clearer, deeper and more practical characterization of consciousness, for natural (biological) and artificial (machine) systems. All animals have APFGP but traditional AI systems do not. This work reports a new kind of AI systems-conscious machines. Instead of arguing what static tasks a conscious machine should be able to do, this work suggests that APFGP is a computationally clearer and necessary criterion for us to dynamically judge whether a system can become maturely conscious through lifelong development, even if it (e.g., a fruit fly) does not have a full array of primate like capabilities such as vision, audition, and natural language understanding. The results here involve a series of new concepts and experimental studies for vision, audition, and natural languages with new developmental capabilities that are not present in many published systems, e.g., IBM Deep Blue, IBM Watson, AlphaGo, AlphaFold and other traditional AI systems and intelligent robots.

show abstract

Autonomous Programming for General Purposes: Theory and Experiments

Cited by 3 publications

References 30 publications

Conscious Intelligence Requires Lifelong Autonomous Programming For General Purposes

Conscious Intelligence Requires Lifelong Autonomous Programming For General Purposes

Perspective Chapter: Deep Learning Misconduct and How Conscious Learning Avoids It

Machines Develop Consciousness Through Autonomous Programming for General Purposes (APFGP)

Contact Info

Product

Resources

About