AlteregoNets: a way to human augmentation

TL;DR

The paper addresses the challenge of algorithmically simulating internal human representations of perceived objects as flows of thoughts and feelings, rather than static property lists. It introduces AlterEgo nets (per-person neural frameworks) and the BiWheel iterative scheme that generates two interconnected textual streams, the PAS ($$PAS=\{w_i\}$$) and the OAS, through forward and backward iterations to model the dialogue between a person and an object. Key contributions include formalizing the per-person AE net, presenting a simple universal implementation, extending the model to multiobject and personalized nets, and outlining diverse applications such as per-user content generation and crowd perception modeling. This framework advances human augmentation by producing individualized, narrative-like representations of objects that can tailor content and interactions to specific users or groups. Potential impact spans personalized advertising, safety screening, and creative content generation, with future work targeting cross-modal perception and cross-domain synthesis.

Abstract

A person dependent network, called an AlterEgo net, is proposed for development. The networks are created per person. It receives at input an object descriptions and outputs a simulation of the internal person's representation of the objects. The network generates a textual stream resembling the narrative stream of consciousness depicting multitudinous thoughts and feelings related to a perceived object. In this way, the object is described not by a 'static' set of its properties, like a dictionary, but by the stream of words and word combinations referring to the object. The network simulates a person's dialogue with a representation of the object. It is based on an introduced algorithmic scheme, where perception is modeled by two interacting iterative cycles, reminding one respectively the forward and backward propagation executed at training convolution neural networks. The 'forward' iterations generate a stream representing the 'internal world' of a human. The 'backward' iterations generate a stream representing an internal representation of the object. People perceive the world differently. Tuning AlterEgo nets to a specific person or group of persons, will allow simulation of their thoughts and feelings. Thereby these nets is potentially a new human augmentation technology for various applications.

Figures (10)

• Figure 1: Relations between representations of perceived objects. (A): The image commonDQ perceived by Don Quixote in the episode with windmills DQText. (B): The flow of thoughts streamP modeled in this paper. (C): The percept Percept of the input image. (D), (E): The textual and image media representing the human representations. The flow of thoughts is naturally represented as a text, and the percept as an image. (D) depicts the representations constructed in this article. The dashed arrow (AB) denotes partial forming of the mental state bypassing the percept. For example, the feel of a color of a shoePercept may be obtained without forming the whole image of the shoe. Green dashed arrow (AD) denotes optional use of the tools for image annotations 3dparty.
• Figure 2: Illustration to BiWheel scheme (best viewed in color). The support and interleaving streams, are depicted respectively by green and yellow arrows. (A): Perception of the advertisement image (Sect. \ref{['SECTPAS']}). (B): Perception of the historical emblem hammer (Sect. \ref{['SECTPAS']}). (C): Perception of the social scene shown at Fig. \ref{['fig:FIG14C']}. Brown arrows at (A), (B): the elements inserted to the PAS (resp. OAS) stream may be obtained using external tools for image annotations.
• Figure 3: The image from Fig. \ref{['fig:BiWheel']} (C)vaada.
• Figure 4: The images refereed to in Table \ref{['cocacolatable']}: (A), (B), and (C) are from Coca_wanna_drink, Coca_accomplishment, and Coca_aesthetic respectively.
• Figure 5: (Best viewed in color.) (A): The PAS stream. The elements of the support substream are depicted in green, of the interleaving stream in yellow. (B): The OAS stream. The elements of the support substream are depicted in yellow, of the interleaving stream in brown. Optionally, the elements of the interleaving streams may be obtained by using annotation tools. (A) and (B): In the both PAS and OAS streams the elements obtained by using selective summarization are depicted in violet.