Semiotics Networks Representing Perceptual Inference

David Kupeev,Eyal Nitcany
2024-07-02
Abstract:Every day, humans perceive objects and communicate these perceptions through various channels. In this paper, we present a computational model designed to track and simulate the perception of objects, as well as their representations as conveyed in communication. We delineate two fundamental components of our internal representation, termed "observed" and "seen", which we correlate with established concepts in computer vision, namely encoding and decoding. These components are integrated into semiotic networks, which simulate perceptual inference of object perception and human communication. Our model of object perception by a person allows us to define object perception by {\em a network}. We demonstrate this with an example of an image baseline classifier by constructing a new network that includes the baseline classifier and an additional layer. This layer produces the images "perceived" by the entire network, transforming it into a perceptualized image classifier. This facilitates visualization of the acquired network. Within our network, the image representations become more efficient for classification tasks when they are assembled and randomized. In our experiments, the perceptualized network outperformed the baseline classifier on MNIST training databases consisting of a restricted number of images. Our model is not limited to persons and can be applied to any system featuring a loop involving the processing from "internal" to "external" representations.
Artificial Intelligence,Computer Vision and Pattern Recognition,Social and Information Networks
What problem does this paper attempt to address?
The paper aims to address the problem of how to model and simulate the process of human perception of objects and how this perception is conveyed in interpersonal communication. Specifically, the authors propose a computational model to track and simulate people's perception of objects and how these perceptions are expressed through various communication channels (such as text, sound, images, etc.). The main contributions of the paper include: 1. Proposing a new model that combines object perception with the resulting interpersonal communication for simulation. 2. Using a series of personalized images of "observed" and "seen" to simulate the human perception process, providing practical interpretability for understanding network states. 3. Extending the model to any system with cyclic processing from "internal" to "external" representation, not just humans. 4. Modeling the "observed-seen" process through encoder and decoder operations, allowing the behavior of object perception to be represented as a series of iterative processes, and these iterative processes converge to attractors. 5. Developing an attractor-based classifier for traditional computer vision classification tasks, which has good visualization properties and outperforms standard baseline classifiers on small training datasets. 6. Describing several semiotic phenomena regarding individual-object communication and interpersonal communication. The paper also introduces some new concepts, such as Consciousness Networks (CONNs), which is a model capable of simulating individual perception of objects and awareness in interpersonal dialogue. Additionally, the paper introduces the concept of dual orbits as fixed points in interpersonal communication. In summary, this paper mainly focuses on establishing a theoretical framework for better understanding and simulating the complex processes of human perception and communication, and attempts to apply it to computer vision tasks.