Cross-domain and within-domain synaptic maintenance for autonomous development of visual areas

Abstract

Abstract-Where-What Networks (WWNs) is a series of developmental networks for the recognition and attention of complex visual scenes. One of the most critical challenges of autonomous development is task non-specificity, namely, the network is meant to learn a variety of open-ended task skills without pre-defined tasks. Then how does a brain-like network develop skills for object relation that can generalize using implicit symbol-like rules? A preliminary scheme of uniform synaptic maintenance, which works across a neuron's sensory and motor domains, has been proposed in our WWN-9. In the new work here, we show that cross-domain and within-domain synaptic maintenance gains superior generalization than using the uniform synaptic maintenance scheme. This generalization enables the WWN to automatically discover symbol-like but implicit rules -detecting object groups from new combinations of object locations that were never observed. By "symbol-like but implicit rules", we mean that the development program has no symbols and explicit rules, but symbol-like concepts (location, type) and implicit rule (two specific type objects must present concurrently -group) emerge as the firing patterns of the motor area and are used by the control. Moreover, the process of synaptic maintenance corresponds to the genesis (and adaptation) of cell connections and our model autonomously develops the Y area into two subarea, early area and later area, in charge of pattern recognition and symbolic reasoning respectively.

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

“…Cresceptron did not deal with time. A developmental approach that deals with both space and time in an unfired fashion using a neural network started from Developmental Networks (DNs) [47] whose experimental embodiments range from Where-What Networks 1 (WWN-1) [48] to WWN-9 [49]. The DNs went beyond vision problems to attack general AI problems including vision, audition, and natural language acquisition as emergent Turing machines [41].…”

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

“…Higher pains (e.g., loss of loved ones and jealousy) and higher pleasure (e.g., praises and respects) develop at later ages from lower pains and pleasures, respectively. (b) Synaptic maintenance-grow and trim the spines of synapses [2,19]-to segment object/event and motivate curiosity. Each synapse incrementally estimates the average error β between the pre-synaptic signal and the synaptic conductance (weight), represented by a kind of neural transmitter (e.g., acetylcholine [33]).…”

A Model for Auto-Programming for General Purposes

“…(b) Synaptic maintenance -grow and trim the spines of synapses [15], [16] -to segment object/event and motivate curiosity. Each synapse incrementally estimates the average error β between the pre-synaptic signal and the synaptic conductance (weight), represented by a kind of neural transmitter (e.g., acetylcholine [109]).…”

“…The X and Z areas are supervised by physics, including self, teachers, and other physical events. Through the synaptic maintenance [15], [16], some Y neurons gradually lost their early connections (dashed lines) with X (Z) areas and become "later" (early) Y areas. In the (later) Parietal and Temporal lobes, some neurons further gradually lost their connections with the (early) Occipital area and become rule-like neurons.…”