The Mobile Advanced Artificial Intelligence Neural Network Project:
This project will result in an application that will run on mobile devices (iPhone/iPad/Android) which will use advanced artificial intelligence (AI) techniques that will perform machine learning tasks, adaptive pattern recognition and dynamic optimization 24/7 in the cloud. This first version of the free mobile app will result in a functioning proof of concept which can be further developed and driven by recommendations from users regarding features to be added. The next phase would be to perform crowd-sourced (mobile) computing and update the experiential data with the collective intelligence in the cloud, then update the local mobile version with the latest version of the Neural Network (NN). We envision that this project will evolve into an advanced AI capable of demonstrating imagination and independent thought, which will learn through feedback and experiences with users.
Although it is not the focus of this project, we have added an element of augmented reality in order to enhance elements of fun and interactivity between the ai personality and the user in the initial free app that will be available in the mobile app stores. The ai personality can "see" what is in view of the camera and, after sufficient training through experience from multiple users, can recognize and interact with items in the scene in a manner similar to that of a human child. As experience grows, the intelligence continues to learn new capabilities and skills, just as a human. Given a set of pre-defined interaction rules, a future version of this intelligence could be placed in an interactive robot or toy. (No, it won't take over the world!)
You may think that this project appears too good to be true. However, I have put much thought into how these things can be accomplished and have broken-down all the necessary features into a long sequence of development tasks that results in a detailed explanation of how to successfully accomplish these tasks.
This project is a huge undertaking, but much of the groundwork has already been proven. Ideas for creating a truly advanced artificial intelligence have existed for decades. Capabilities of computers have only fairly recently reached the level where the average user can own a mobile device with sufficient rocessing power. Additionally, I have been involved in much of the applied research done in this software development space. I have come up with a method of breaking-down a scene (augmented reality) into recognizable sub-components, many of which may be many parts related to a whole. An example would be the separate blocks of wood which are parts of a chair, belonging to a group of furniture, which belongs to the kitchen, a room which belong to the home, etc. Initially the simulated intelligence developed from this project is much like that of a child, which is very capable of learning but will benefit from a jump-start with some basic knowledge in terms of concepts and categorized items in which to relate with each other. For instance, ask an artificial intelligence "show me an image of what you imagine a cat to look like." What would you expect to see? How would Siri react? The process I plan on integrating into this project will result in an intelligence capable of artistic representation based-on collective experience.
Coming soon: a video explaining the concept!
Ref: A Self-Learning Evolutionary Chess Program
Ref: Kurzwell (Ray) Singularity - Who am I? - What have I become?
Status: I am seeking funding for this project.