Featured

Self-Motivated Intelligent Agents

1. Introduction 1.1 Agents and Environments An agent-based system is comprised of two main functions, the agent function A(s) and the environment function E(a), where s is an observable state of the environment and a is an action selected by the agent. The input received by each function at any given time is the output … Continue reading Self-Motivated Intelligent Agents

Advertisements
Featured

Adaptive Template Model of Intelligence

Click to see this project on GitHub 1. Introduction 1.1) Problem Overview Traditionally, template-matching algorithms have been used for things like digital image processing and visual pattern recognition. There are typically sets of small, two-dimensional filters that are moved across a gray-scale image in order to detect instances of low-level visual patterns. Pattern recognition via … Continue reading Adaptive Template Model of Intelligence

Higher Cognition of Intelligent Agents

Visual analogy The human brain takes advantage of a spatial system that originally developed for visual cognition and applies higher cognitive functioning to map concepts onto a mental space in the form of a visual analogy. Non-spatial relations between ideas are assigned to connections, replacing their use as an indicator of spatial distance with a […]

Constraint-based Hierarchical Pattern Detection

Data is received by the constraint-based hierarchical detection system (CHD) in the form of a two-dimensional grid, where each cell contains a single object that holds information. A system is made of multiple grids, each stacked on top of the previous. Between each pair of adjacent grids lies a system called a mapper. A mapper … Continue reading Constraint-based Hierarchical Pattern Detection

Generative Logical Systems

Project on GitHub 1. Objects and Storage A data store has a capacity which limits the amount of information that it can hold. When the capacity of a store is met, the contents are mapped by a function M(S) to another structure called an object. The store is then free to hold new new information … Continue reading Generative Logical Systems

Linguistic Frame Learning

Linguistic frame learning  is an approach to natural language in machine learning, using a frame-based knowledge accumulated over time through detection and generalization of rules from a set of observations (i.e. examples of natural language), in order to define a given language in terms of a constraint satisfaction problem. 1. Language Learning 1.1 Layers of Knowledge Language … Continue reading Linguistic Frame Learning

Deep Representation of Semantics in Natural Language

1. Linguistic Units 1.1) Nouns and Verbs The building blocks of any language are primarily two linguistic units called nouns and verbs. A noun is typically defined as a ‘person, place, or thing’, however this definition fails to account for some of the more abstract concepts which still fall under the category of nouns, like … Continue reading Deep Representation of Semantics in Natural Language

Logical Foundations of Knowledge

Function Words The precise meaning of a word is dependent on contextual knowledge. For instance, the function word “with” can either refer to an instrument or a co-agent. One being an object and the other an agent, the implications differ depending on the context in which the term is used. Until it can be determined … Continue reading Logical Foundations of Knowledge

Mathematical Foundations of Knowledge

Overview As intelligent beings, we’re able to interpret constraints that effect our understanding of the spatial, temporal, social, political, and economic world that we inhabit. Our intuitive realization of boundaries, I argue, stems from the built-in neurological mechanism of classification and pattern recognition. In order to frame experiences in a meaningful way, there must be … Continue reading Mathematical Foundations of Knowledge

On Relational Knowledge Representation

Overview In order to effectively represent knowledge so that it may be utilized in future problem-solving, it is necessary to encode relational information beginning at the spatial-temporal dimension and working upward with increasing complexity, such that the relations at higher levels are constructed from those below, with the result being a theoretically infinite hierarchy of … Continue reading On Relational Knowledge Representation