Where-what network 1: “Where” and “what” assist each other through top-down connections

Abstract

This paper describes the design of a single learning network that integrates both object location ("where") and object type ("what"), from images of learned objects in natural complex backgrounds. The in-place learning algorithm is used to develop the internal representation (including synaptic bottomup and top-down weights of every neuron) in the network, such that every neuron is responsible for the learning of its own signal processing characteristics within its connected network environment, through interactions with other neurons in the same layer. In contrast with the previous fully connected MILN [13], the cells in each layer are locally connected in the network. Local analysis is achieved through multi-scale receptive fields, with increasing sizes of perception from earlier to later layers. The results of the experiments showed how one type of information ("where" or "what") assists the network to suppress irrelevant information from background (from "where") or irrelevant object information (from "what"), so as to give the required missing information ("where" or "what") in the motor output.

Brains as naturally emerging turing machines

“…The DN has had several versions of experimental embodiments, called Where-What Networks (WWNs), from WWN-1 [14] to WWN-7 [43]. Each WWN has multiple areas in the Z areas, representing the location concept (Location Motor, LM), type concept (Type Motor, TM), or scale concept (Scale Motor, SM), and so on.…”

“…In addition to the similarity to Hopfield Network [13] and LSTM [11] cited therein, Graves et al 2014 [9] appeared to have a series of mechanism similarities with DN 2011 [30] along with DN embodiments Where-What Networks, WWN-1 2008 [14] to WWN-7 2013 [43]. To facilitate understanding their conceptual relation, let us see some of the similarities with correspondence of concepts: (1) The finite state machine (i.e.…”

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

“…Where-What Networks (WWNs) [8] first introduced in 2008 is a biologically plausible developmental model designed to integrate the object recognition and attention namely, "what" and "where" information in the ventral stream and dorsal stream respectively. Therefore, multiple concepts (e.g., type, location, scale) can be learned concurrently in the network.…”

Skull-closed autonomous development: WWN-6 using natural video

“…So far, WWN has six versions. WWN-1 [11] can realize object recognition in complex backgrounds performing in two different selective attention modes: the top-down positionbased mode finds a particular object given the location information; the top-down object-based mode finds the location of the object given the type. But only 5 locations were tested.…”