Paper ID: 2310.05212

Interpretable Semiotics Networks Representing Awareness

David Kupeev, Eyal Nitcany

Humans perceive objects daily and communicate their perceptions using various channels. Here, we describe a computational model that tracks and simulates objects' perception and their representations as they are conveyed in communication. We describe two key components of our internal representation ("observed" and "seen") and relate them to familiar computer vision notions (encoding and decoding). These elements are joined together to form semiotics networks, which simulate awareness in object perception and human communication. Nowadays, most neural networks are uninterpretable. On the other hand, our model overcomes this limitation. The experiments demonstrates the visibility of the model. Our model of object perception by a person allows us to define object perception by 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. Within our network, the internal 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.

Submitted: Oct 8, 2023